Design a site like this with WordPress.com
Get started

What’s the tech behind ChatGPT and the risks for users and society? An educator’s perspective

In this video interview for Faculti.net, I explain what’s the technology behind ChatGPT and what are its risks for users and society at large.

Background and context of the research

In April 2021, the EU presented the Draft AI Act, which if adopted will be the first legislation attempting to regulate AI. Its key tenet is that certain AI systems are to be regarded as high risk, e.g. when AI is used to decide who can access to education and how they are assessed. However, the AI Act was designed with narrow / applied AI, and even though the European Parliament last week has introduced some amendments to account for the rise of foundation models, it’s unlikely that few tweaks are going to be enough to minimise the risks of these systems.

In October 2022, the European Commission’s Expert Group in AI in Education (of which I’m a lucky member) published the EU Ethical Guidelines on AI in Education, to try and help teachers and students to embrace AI in a responsible way. Similar to the AI Act, the Guidelines were conceived before a certain fancy chatbot became viral…

In November 2022, OpenAI’ ChatGPT exploded:

  • ChatGPT crossed 1 million users in 5 days of launch
  • 100 million monthly active users in January (2 months after launch): fastest-growing consumer application in history (TikTok 9 months!) (UBS 2023)
  • Integrated in search engines, Outlook, Word, etc.: ubiquitous AI

That’s why I’ve started working on the book The Law of Large Language Models, that will be published for Cambridge University Press as part of the Cambridge Elements in Data Rights and Wrongs, book series edited by Megan Richardson, Damian Clifford, Jeannie Marie Paterson, and Julian Thomas

The technology

This graph is a visualisation of the tech underlying ChatGPT. I find it useful as often these terms are mistakenly used interchangeably.

Foundation models

A foundation model is “any model that is trained on broad data (generally using self-supervision at scale) that can be adapted (e.g., fine-tuned) to a wide range of downstream tasks” (Stanford Center for Research on Foundation Models 2021)

The definition can be broken down into some key features, as helpfully explained by Babar M Bhatti (2023):

Pretrained – it’s not like when you ask something to ChatGPT, ChatGPT goes on the internet to find the answer: it’s been pre-trained with huge amounts of data (GPT-3 was trained in September 2021)

Generalised – not specific for a task

Adaptable – through prompting

Large – e.g., GPT-3 has 175B parameters and was trained on about 500,000 million words, equivalent of over 10 lifetimes of humans reading nonstop! (Bhatti 2023)

Self-supervised — no specific labels are provided and the model has to learn from the patterns in the data which is provided

There are two aspects of foundation models that make me worry:

Emergence – unpredicted abilities that are not present in smaller-scale models (Wei et al. 2022)

Homogenisation the same method is used in many domains, which allows for powerful advances but defects of the foundation model are inherited by all the adapted models downstream (Bommasani et al. 2021)

Generative AI

There are some foundation models that are not about generating content (text, speech, etc.). Key non-generative tasks include prediction and classification.

The foundation models who are making the headlines at the moment are those that fall within the scope of generative AI. The latter is a type of foundation model that generates images, music, speech, code, video or text, while it interprets and manipulates pre-existing data.

Examples of generative AI include:

Text-to-music generation e.g. Google’s MusicLM

Text-to-image e.g. Dall-E2, Crayon, Stability

Text-to-video e.g. Meta’s Make-me-a-Video and Google’s Imagen Video

My focus in on Large Language Models (LLMs) and text-to-text generative AI.

Large Language Models (LLMs)

Large Language Models (LLMs) are a type of generative AI typically using natural language processing (NLP) to generate text.

NLP is the “field covering knowledge and techniques involved in the processing of linguistic data by a computer” (ISO 24613-1:2019(en), 3.9)

In other words, LLMs are a type of machine learning model that can perform a variety of NLP tasks e.g., generating and classifying text, answering questions in a conversational manner and translating text (Rouse 2023)

Trained with immense amounts of data and use self-supervised learning to predict the next ‘token’ (typically a word) in a sentence, given the surrounding context. The process repeated until it’s accurate.

Increasing the scale of language models (e.g. model parameters) can lead to better performance and efficiency on a range of downstream tasks (Devlin et al., 2019; Brown et al., 2020), as well as the emergence of unpredictable abilities (Wei et al. 2022)

The best-known LLM tool is ChatGPT, from OpenAI, a chatbot that uses AI and statistics to predict the next word in a sentence based on the preceding words (other LLMs are BERT and T5)

Is AI inferior to humans for cognitive tasks?

The conventional wisdom is that while AI is better than humans at data-driven decision-making tasks, it is still inferior to humans for cognitive and creative ones.

We can no longer take for granted. Here some examples that make me worry:

  • A judge in Colombia used ChatGPT to assist in a court ruling
  • A judge in India used it to decide on bail
  • “AI Drake” song – a made-up song, made to sound as if Drake were singing it – has gone viral (and temporarily hosted on Spotify)
  • Grimes is authorising anyone to use her voice for AI-generated music, splitting 50% of royalties
  • GPT-4 has convinced a TaskRabbit worker to solve a Captcha puzzle by claiming to be a visually impaired person

All this in the last couple of months.

ChatGPT: societal risks

Bias, discrimination, stereotyping

•Threat to jobs – law teachers and judges among the most exposed (Eloundou et al. 2023)

New slavery: 150 African workers behind ChatGPT (TikTok and Facebook) voted to unionise at a landmark meeting in Nairobi

Sustainability – carbon footprint = 175,234 Danes (Ludwigsen 2023)

Educational inequity: GPT-4 safer and more accurate but $20 a month (Bjork 2023)

•ChatGPT passes the bar exam (Katz et al. 2023): potential for cheating? Are we using the right mechanisms to select the lawyers of the future?

ChatGPT: individual risks

•Plagiarism and academic integrity

•Misinformation / fake news – ChatGPT hallucinates

•Defamation and hate speech – A mayor in Australia is suing OpenAI for defamation over inaccurate information about him provided by ChatGPT

•Mental health – a young Belgian committed suicide following weeks of conversations with an AI-driven chatbot named Eliza (Lovens 2023)

Dubious content practices – copyright

Dubious data practices – privacy and data protection

ChatGPT: the benefits in education

•Generate list of topics (for essay, research paper, etc.)

Proofread

Paraphrase e.g. to provide accessible feedback but “To avoid copyright concerns, teachers should paste their own feedback into AI platforms, rather than students’ work” (Eager 2023)

Translate

•Teach students new skills e.g. how to write meaningful prompts

91% of companies that are hiring want to bring on employees with ChatGPT experience (Resume Builder 2023)

Assessment – my experience

•I’ve asked ChatGPT: ‘How has the internet impacted Media Law? Please write a fully referenced essay as if you were a law student.’

•This is what ChatGPT answered: ‘…’

Critically reflect on the ChatGPT’s answer to my prompt. What is missing? What are its limitations? Can you find the adequate sources to back what ChatGPT is saying? Please read this module’s AI Policy here before working on the assessment”

•Lack of critical analysis

Referencing – made up and not following official styles

Out of date data (pre-trained – September 2021)

Incorrect information

Conclusions

•ChatGPT presents serious individual and societal issues

•It can be a useful tool for students – emphasis on critical and responsible use of AI: teach students not to trust it!

•Risk that illegal content generated by ChatGPT exposes teachers and students to liability

•A shift toward more honest conversations with students and authentic assessment

Is patent law conducive to sustainability in an AI world?

With my co-authors – the fab Gabriele Cifrodelli (Glasgow) and Artha Dermawan (Max Planck & Lapland) – we’ve been researching the question of whether European patent law is sustainable in an AI world.

Our attempt to answer can be found in the chapter “Sustainable Patent Governance of Artificial Intelligence: Recalibrating the European Patent System to Foster Innovation (SDG 9)” which will form part of the Elgar Companion on Intellectual Property and Sustainable Development Goals, a forthcoming book edited by Bita Amani, Caroline Ncube, and Matthew Rimmer for Edward Elgar.

This chapter reflects on what a sustainable patent governance looks like in the age of the rise of inventive AI. To do so, we critically evaluate the relationship between Artificial Intelligence, European patent law, and sustainability with a focus on the Sustainable Development Goal 9, i.e. to build resilient infrastructure, to promote inclusive and sustainable industrialisation, and to foster innovation. In particular, we consider SDG targets 9.1, 9.2, and 9.5 for their emphasis on equitable access to innovation, inclusive industrialisation, and an imperative to consider the national circumstances of developing countries.

While the contribution of patent law to SDG 9 may prima facie appear as straightforward – both are all about promoting innovation – we problematise this relationship by shedding light on the unsustainability of patents.

In Section 1, we analyse patent data related to climate change mitigation technologies to discover that, under the European Patent Convention, there has been a significant growth in AI-related patenting, but that developing countries are being left behind.

In Section 2, to overcome the decreased access to innovation stemming from an inventive step requirement that does not match the reality of AI-powered ingenuity, we focus on AI datasets and suggest a recalibration that revolves around the concept of Therapeutics Data Commons.

In Section 3, we suggest a more nuanced understanding of AI inventorship to include Global South perspectives.

We put forward that to make tangible progress towards SDG 9, patent law ought to be part of a movement of radical change. As our analysis of patent data in Europe shows, AI-related patents that can be used for sustainability purposes are on the
rise. However, SDG 9 sees the encouragement of innovation as a way of tackling wider societal challenges, and doing so in a way that is fair, equitable, and mindful of national circumstances, especially in developing countries.

Sustainable AI requires sustainable governance of patents. AI sustainability requires a change in the patent system and re-appreciation of some of its core concepts. This should include the rethinking of innovation as an instrumental good, a recalibration of the inventive step to ensure equitable access to AI-generated innovation, and the acknowledgement that doctrines that may be groundless in
developed countries (eg AI inventorship) may well be justified in the Global South.

The march towards progress should be paused if we cannot ensure that innovation becomes fairer and more inclusive. Patent law should be part of this progress, lest it become just another residue of an unjust past.

The preprint is available on SSRN.

Funded PhD in the just transition to AI – Just AI Lab (University of Stirling)

Exciting funded PhD opportunities

Do you want to contribute to the just transition to socially just, trustworthy, responsible, open, inclusive and sustainable AI? Come do a funded PhD with us at the Just AI Lab of the University of Stirling!

Thanks to this £3.6m investment, we can offer a number of 3-year full-time equivalent PhD studentships on a fully funded basis, to both UK and international applicants!

We welcome expressions of interest from all candidates who meet the University’s entry criteria for Postgraduate Research degrees. We especially welcome expressions of interest from people from groups that have been under-represented in the UK Post Graduate Research community in the past, including people from ethnic minorities, women, disabled people, and people from care-experienced or socio-economically disadvantaged backgrounds.

The Just AI Lab

Funded PhD students will join the Just AI Lab, a collaborative, inter-disciplinary cluster of researchers from across the University of Stirling, Scotland (UK).

PhD students will be working on complementary doctoral projects about the just transition to Artificial Intelligence (AI).

AI – including autonomous systems, machine learning, and other data-driven smart systems – is increasingly taking on crucial roles in many contexts, including health and wellbeing, tourism and leisure, security, law, human resources and education. Often, the humans who interact with and use these systems have a limited understanding of how they work, and consequently how they arrive at the judgements and decisions they output, with important consequences for trust in outputs and understandings of uncertainty. 

Against this backdrop, the core purpose of the Just AI Lab is the just transition to socially just, trustworthy, responsible, open, inclusive and sustainable AI through impactful mission-oriented interdisciplinary research. For the transition to be just, we must understand, use, and engage critically with AI in all aspects of human endeavour. Through Just AI, we embrace the vision of the Scottish AI Strategy of a fairer, greener, more prosperous, and outward looking Scotland. Our key areas of interest are: Just Wellbeing, Just Work, Just Services, Just Commerce and Just Governance.

Application domains

Proposals are expected to focus on the just transition to AI with regard to one of the following application domains.

Just Governance. When it comes to governing AI, Scotland’s AI Strategy refers to the OECD principles for the responsible stewardship of trustworthy AI, including the need to design AI in a way that respects human rights, democratic values, diversity, transparency, safety, and accountability. How can we adapt traditional product safety frameworks (e.g. CE certification) to AI systems to ensure data safety? The UK Government is departing from the GDPR, how can we make sure that the legal mechanisms around transparency and fairness in algorithmic decision-making are not watered down? The EU has proposed new AI-specific legislation; what lessons can be learned in the UK? How can law work with computer science on the development of AI tools for the functional analysis of law, especially in the context of alternativ dispute resolution? Can the human and societal risks of facial recognition in policing be minimised? How can we overcome privata data enclosures in order to re-use data for the common good?

Just Wellbeing. AI will be pivotal to Scotland’s ambition to ensure a high quality of life for the people of Scotland, with a focus on prevention and early intervention (Programme for Government). AI is increasingly used, and increasingly blurs the lines between leisure, games, sport, and health. E.g., commercially available smart speakers, used for leisure, also have applications in the care sector. What is the influence of these smart speakers on older people’s lives? What are the barriers? How do they influence quality of life? Are there concerns about privacy? How can duties of care be coded? How are assistive technologies used for healthy ageing, especially in relation to sport reminiscence and sporting heritage? AI is becoming precious to improve health outcomes by changing behaviour. Are there groups who would benefit most from using AI to manage health outcomes e.g., those in rural areas?

Just Work. While there are growing calls within Europe to address the risks and potential negative impacts AI systems may have on work, dominant narratives position AI as offering assumedly straightforward solutions to complex work issues. Under the Scottish Programme for Government, technological innovation is seen ‘as a key driver for improved productivity and work standards/conditions.’ Such narratives do not address how workers in AI-mediated work settings employ their professional expertise and capacities for agency and judgement (alongside automated and assisted decision-making systems) critical for the delivery of essential services in society. How are and can alternative AI-Work narratives be co-created and mobilized? Will the creative industries be disrupted by AI-generated art, music, literature, and film? Are we witnessing the full realisation of the Adorno/Horkheimer notion of work and leisure time coming to increasingly resemble one another?

Just Services. AI is increasingly used in the provision of essential services such as education and healthcare, by public and private actors as well as hybrid organisms (e.g. public-private partnerships). E.g., pursuant to the Scottish Programme for Government, we are moving towards digitally-enabled schools and campuses as well as other sites of lifelong learning and what kinds of critical AI-Data literacies are required to develop stronger public understandings of AI and how to facilitate discussion of challenges and opportunties? How can we ensure a just transition to AI in education in a market dominated by commercial platforms? Have schools and universities the resources and skills to assess the threat AI poses to both teachers and students? Key questions apply also to other sectors.  What public interest safeguards are needed in AI procurement by public bodies? What are the rule of law implication of the use of AI in the public sector settings and in public private partnership settings e.g. in the transport and law enforcement sectors?

Just Commerce. Business uses of AI abound, ranging from chatbots used in e-commerce to quantum AI in the FinTech sector. AI innovation is a primary area of economic growth in Scotland, where Government is committed to ‘(u)se our trade and investment plans to set the direction on new market opportunities, attracting the high quality investment and technologies needed.’ To ensure the just transition to AI in commerce, a number of questions remain unanswered. What is the environmental impact of AI adoption, especially by SMEs and microenterprises that do not have the resources to offset the relevant externalities? Can AI save the high-street or will it be its end? How can we ensure that companies use AI humanely? How do we prevent data-driven personalisation from becoming consumer manipulation or even discrimination?

People

The Just AI Lab is directed by Associate Professor Guido Noto La Diega (Law and Philosophy) with Co-Directors Associate Professor Greg Singh (Communications, Media & Culture), Dr Terrie Lynn Thompson (Education), Professor William Webster (Management, Work and Organisation), Professor Anna C. Whittaker (Sport), and Dr Sandy Brownlee (Computing Science & Mathematics).

Other potential supervisors associated to the Lab are: Professor Kevin Grant, Professor Iain MacRury, Dr Keiller Nogueira, Professor Hon-Lin Yu, Professor Richard Haynes, Dr Hannah Graham, Dr Mo Egan, Dr Lorna Gibb, Dr Tom Kane, Dr Jenni Connelly, Dr Naja Yusof, Dr Diana Miranda, Professor Mike Wheeler, Dr Carolyn Wilson-Nash, Dr Zoi Krokida, Dr Melissa Avdeef, Dr Fay Niker, Dr Wasim Ahmed, Dr Simona Hapca, Dr Michaela Hruskova, Professor Jayne Donaldson, Professor Ashley Shepherd, Dr Seda Erdem, Dr Ellie McDonald, Dr Victoria Esteves, Dr Sarah Thomson, and Dr Ailsa E Millen.

Example projects

The example projects below are intended to inspire thought and provide context, but you are expected to submit an original project proposal that aligns with the research cluster.

How can the regulation of AI and AI-generated decisions / outputs ensure the just transition to AI and its use to tackle societal challenges e.g. climate justice, social justice, etc.?

How to deal with uncertainty, trust, and ethical design principles for transparent automated decision-making within selected domains such as organisational behaviour, governance, administration?

How to develop and test AI tools to enhance healthy life choices for psychological and physical wellbeing?

Guidance and key dates

The Institute of Advanced Studies (IAS) will be holding two webinars to provide guidance and answer questions that prospective postgraduate researchers may have, these will take place on 28 February and 9 March. Register for a webinar.

Submit your expression of interest by 14th April. Select Novel methodologies to support future research and better informed policy and then ‘Just AI Lab’.

Once the deadline for expression of interest has passed, expressions of interest will be assessed by clusters and nominations for an award made to the award panel. It is expected that offers will be made by 1 June 2023.

By 15th June, following the offer of an award a formal application for a PhD programme should be made, the process for applying for a PhD programme at Stirling can be found here. In order to receive an award, you must evidence that you meet eligibility criteria and secure an unconditional offer in advance of the start date of the PhD programme, specifically, 1 October 2023.

Before applying, you are encouraged to reach out to a potential supervisor (one of the researchers listed under ‘People’ in the previous section.

Make sure to read the full guidance here.

Express your interest

New chapter! “Ending Smart Data Enclosures: The European Approach to the Regulation of the Internet of Things between Access and Intellectual Property”

As Europe convincingly embraces quantum computing – with its potential to solve complex societal problems ranging from health through energy transition to climate justice – there can no longer be doubt that the big data generated by IoT systems must become accessible and smart data enclosures must end. While on the face of it the most recent legislative interventions go in the direction of more open, accessible, and re-usable data, they tend to balance (or even neutralise) that ethos with provisions that IoT big tech can use to retain and strenghten data enclosures.

After becoming a global leader in the field of digital regulation with the General Data Protection Regulation (GDPR) – which prompted similar laws across the globe, from California through Turkey to China – the European Union (EU) is at it again with the Artificial Intelligence (AI) Act. The latter will ban certain practices (e.g. face recognition for law enforcement purposes) and impose human oversight, transparency, and other requirements on providers of high-risk AI systems (e.g. automated systems used in recruitment) . When it comes to the IoT, conversely, no ‘IoT Act’ is on the horizon. This makes it harder to get a clear sense of how this socio-technological phenomenon is being regulated in Europe.

While in the future we expect to see IoT-related interventions in the fields of standard essential patents (SEPs), standards, product liability, and common European data spaces, there are already some instruments that aim to introduce tools to access otherwise locked-up smart data. In this chapter – that will be included in Stacy-Ann Elvy and Nancy Kim (eds), The Cambridge Handbook on Emerging Issues at the Intersection of Commercial Law and Technology (Cambridge University Press 2023) – I focus on the relevant provisions in the Open Data Directive, Data Governance Act, the Digital Markets Act, the Digital Services Act, and the AI Act, and see how data access and transparency obligations are balanced against IPRs. Previous research looked at how this balance is struck under the GDPR and the proposed Data Act. Those studies shed light on the emerging conflict between an approach to data governance informed by human rights, sustainability, and social justice – whereby citizens must be given data control, which includes the right to access, reuse, and port the data – and a market-oriented neoliberal one that regards data solely as a commodity to be appropriated and monetised. This paper tests whether those findings can be generalised, and in particular whether it can be said that in the European regulation of smart data the reasons of openness are prevailing over those of private property.

Either the tension between data access and property is holistically addressed (including full data access by end-users, especially in the context of private enforcement), or we will not be able to overcome smart data enclosures, which in turn will have a detrimental effect in terms of potential for smart data to be used for the common good.

My book is out! Internet of Things and the Law: Legal Strategies for Consumer-Centric Smart Technologies

Today is the day! On Friday 14th October 2022, Routledge has published my book Internet of Things and the Law: Legal Strategies for Consumer-Centric Smart Technologies. I will present it next week at the Nexa Center for Internet & Society (discussing with Marco Ricolfi and Jacopo Ciani Sciolla) and in December at the Faculty of Advocates in Edinburgh (an event kindly generously organised by The Royal Society of Arts).

The book has been welcomed as the most comprehensive and up-to-date analysis of the legal issues in the Internet of Things (IoT). For decades, the decreasing importance of tangible wealth and power – and the increasing significance of their disembodied counterparts – has been the subject of much legal research. For some time now, legal scholars have grappled with how laws drafted for tangible property and predigital ‘offline’ technologies can cope with dematerialisation, digitalisation, and the internet.

As dematerialisation continues (see NFTs frenzy), this book aims to illuminate the opposite movement: rematerialisation, namely, the return of data, knowledge, and power within a physical ‘smart’ world. This development frames the book’s central question: can the law steer rematerialisation in a human-centric and socially just direction?

To answer it, the book focuses on the IoT, the sociotechnological phenomenon that is primarily responsible for this shift. After a thorough analysis of how existing laws can be interpreted to empower IoT end users, I leave the reader with the fundamental question of what happens when the law fails us and concludes with a call for collective resistance against ‘smart’ capitalism.

I am grateful to the leading thinkers who endorsed the book. They kindly said:

This book not only advances our understanding of IOT policy but also serves as a model for future work in the law and political economy of technology policy” (Professor Frank Pasquale , Brooklyn Law School, author of the bestseller The Black Box Society)

“Internet of Things and the Law: Legal Strategies for Consumer-Centric Smart Technologies is a thorough exposition of the regulation of the Internet of Things. This is a timely and brilliant addition to scholarship that should inform forward-thinking regulatory approaches” (Professor Caroline B Ncube , Professor and SARChI Research Chair in Intellectual Property, Innovation and Development, University of Cape Town)

A wonderfully informative and deeply reflective study of the Internet of Things from a socio-legal perspective, presented to us by one of the leading experts in the field. The book should be read especially by those interested in how European law might effectively regulate an Internet dominated by ‘Things’ and how people acting collectively can harness their power to reshape the future” (Professor Megan Richardson, Professor of Law, University of Melbourne, and Chief Investigator in the ARC Centre of Excellence for Automated Decision-Making and Society)

The only comprehensive and thorough legal analysis of IoT available as yet, which beautifully combines technological savvy with an admirable love for polemics” (Professor Marco Ricolfi , Co-Director of the Nexa Centre for Internet & Society; Professor of Commercial Law, Università degli Studi di Torino; Equity Partner at Weigmann Studio Legale)

The book has also been positively reviewed by Carlos Muñoz Ferrandis in GRUR International (Oxford), Marc Stuhldreier in The Journal of World Intellectual Property (Wiley); Felipe Romero Moreno in the International Review of Law, Computers & Technology (Routledge); Daria Motovilova and Nataliya Kozlova in Law & Digital Technologies (Russian Academy of Science).

Like all knowledge, a book is a collective endeavour. I wish to thank Northumbria University for granting my sabbatical request, and the University of Stirling for providing me with a generous research allocation. Thank you to the library and administrative staff at both universities for their outstanding professionalism. I am much obliged to Siobhán Poole and everyone at Routledge for believing in this project and being patient while I was missing all the deadlines partly due to deadly viruses, relocations, and job changes.  

I owe a debt of gratitude to my Maestro Luca Nivarra and to all those who selflessly mentored me over these years, namely – in alphabetical order – Ray Arthur, Eva-Marina Bastian, Frances Burton, Enrico Camilleri, Lilian Edwards, Sue Farran, Martin Kretschmer, Dave McArdle, Christopher Millard, Lars Mosesson, Andrew Murray, Marina Nicolosi, Michele Perrino, Fabrizio Piraino, Jenny Preston, Chris Reed, Megan Richardson, Marco Ricolfi, Michael Stockdale, Alain Strowel, Elaine Sutherland, Ian Walden, Jason Whalley, Tony Ward, and Hong-Lin Yu. Thanks for leading by example!

In the years I have worked on this book, my students, PhD students, research assistants, and supervisees have been a source of constant inspiration. They have meantime become successful colleagues. It is impossible to name them all, but I wish to single out Rachel Allsopp, Valentina Borgese, Paolo Burdese, Luca Dell’Atti, Cameron Giles, Zihao Li, Daria Onitiu, Alessia Palladino, Samantha Rasiah, David Sinclair, James Stacey, Pete Tiarks, and Giuseppe Zago. Homines dum docent discunt.

I wish to acknowledge the crucial role played by my research participants. An honourable mention goes to IoT leaders Alexandra Deschamps-Sonsino, Laura James, Joshua Montgomery, Peter Bihr, and Alasdair Davies. Thanks to you I gained a better understanding of the IoT and of the importance of opening it up.

This book would have not been possible without the love and support of my family. Thank you madre and babbo for regarding my weirdness as uniqueness, for teaching me to embrace it and celebrate it. I wish to thank Dr JC Bell, my amazing partner, who stood by me despite my erratic behaviour while working on this book, including waking up in the middle of the night to scribble confused ideas most of which never made it into the final draft, fortunately. I could have not chosen a better (lockdown or otherwise) companion. Thank you for expressing your love through ‘extensive notes.’

Will the Data Act open up machine data and help achieve sustainability?

I am delighted to share the preprint of Guido Noto La Diega & Estelle Derclaye, ‘Opening Up Big Data for Sustainability: What Role for Database Rights in the Fourth Industrial Revolution?’ in Ole-Andreas Rognstad, Taina Pihlajarinne, Jukka Mähönen (eds), Promoting Sustainable Innovation and the Circular Economy: Legal and Economic Aspects (Routledge 2022), available at SSRN

As hundreds of zettabytes of data are being generated by the bio-cyber-physical technologies of the Fourth Industrial Revolution (4IR),[1] it has become untenable to argue that the law needs to introduce incentives to produce more data.[2] Instead, as private corporations amass a wealth of data branded as ‘machine data’ to hide its value, we need incentives to open up enclosed data. The scarcity of resources is the main traditional justification for proprietary approaches.[3] The 4IR heralds an age of abundance of data,[4]  which justifies – even calls for – a commons-inspired open data approach.[5]

Open data is no longer just the mantra of sparse communities of developers and designers. The EU has recently started to embrace open data in a number of recent legislative instruments and proposals to push public entities and citizens to open up their data for the common good. These are epitomised by the new Open Data Directive[6] and the Data Governance Act,[7] respectively. The former creates the conditions for better access to and re-use of public sector information in the belief that openness can benefit society at large.[8]  The latter invites citizens to embrace a philosophy of data altruism on the grounds that better data sharing can not only bring economic value, but also, and more significantly, ‘new ways for tackling societal challenges (e.g. climate change).’[9]

Liberating data from its enclosures is pivotal to the achievement of the UN Sustainable Development Goals (SDG), as we will argue in the next section, but it is apparent that the real guardians of big data – the private corporations that are the key decision-makers in the 4IR – are not doing enough to facilitate the sharing and re-use of data in the public interest, including the pursuit of climate justice. The main instruments recently proposed to curb big tech power do not seem to embrace openness in any meaningful way. This is epitomised by the proposed AI Act. There, market surveillance authorities are granted access to the training, validation and testing datasetsused by the provider of high-risk AI systems, and under certain conditions also to the relevant source code.[10] This is a very limited right to access, which does not allow the re-use of this data and code for the common good, including for sustainability purposes. Crucially, this is consistent with the pro-proprietary stance of the AI Act summed up in the promise that ‘(t)he increased transparency obligations will also not disproportionately affect the right to protection of intellectual property (…) since they will be limited only to the minimum necessary information for individuals to exercise their right to an effective remedy and to the necessary transparency towards supervision and enforcement authorities.’[11] While there may be instances where Intellectual Property (IP) reasons may justify some limitations in the access to and re-use of big data held by corporations, it is our view that, in general, IP should not be used to hinder re-use of data to pursue the SDGs.

As Governments and citizens are being asked to open up data, now it is the time for private companies to open up as well. As the fight for sustainability has never been more pressing and as transparency rises as one of the pillars of global digital constitutionalism,[12] private power must do better and more to allow the access to and re-use of data for the common good. With this in mind, this paper aims to answer the following overarching research question: what does a sustainable legal framework for the achievement of the sustainability goals look like in the 4IR? First, we will illustrate the triple meaning of ‘data sustainability.’ Second, we will critically assess whether the database right (or ‘sui generis right’) can play a role in opening up corporate big data. Third, will imagine how a sustainable framework for sustainable data governance may look like. This focus is justified by the fact that the Database Directive, often accused of creating an unjustified monopoly on data, is in the process of being reformed by the proposed Data Act.

Our key conclusions are:

•Despite a number of laws on open data and access rights still insufficient incentives for private corporations to free data from their enclosures

Data Act is overall a step in the right direction, but it needs to be amended to clarify that

1.All 4IR databases are not protected by the sui generis right;

2.A dichotomy (obtained/created) is not replaced by another dichotomy (produced/inferred) – Risk of data monopolies of derived / inferred data

3.All user rights and freedoms, and ‘openness’ obligations prevail on any contrary contractual and technological measures


[1] By 4IR we mean an umbrella term for those technologies such as Internet of Things (IoT), Artificial Intelligence (AI), blockchain, quantum computing, nanotechnologies, etc. whose increasing fusion and interaction is rewriting the boundaries between biological, physical, and digital (Klaus Schwab, The Fourth Industrial Revolution (Penguin 2017). 4IR is often confused with Industry 4.0, which is more focused on the industrial applications of the IoT (‘smart manufacturing’) see e.g. Christoph Winterhalter, ‘A New Revolution in the Making’ (2018) November-December ISOfocus 3.

[2] Volumeof data/information created, captured, copied, and consumed worldwide from 2010 to 2025’ (Statista, 2022) <https://www.statista.com/statistics/871513/worldwide-data-created/ > accessed 12 January 2022.

[3] Mark A Lemley, ‘IP in a World Without Scarcity’ (2015) 90 NYU Law Review 461.

[4] See Andrea Ottolia and Cristiana Sappa, ‘A Topography of Data Commons: From Regulation to Private Dynamism’ [2021] GRUR International 1. The authors convincingly argue that a data commons approach is key to avoid excessive protection in the data-driven society. Unlike the authors, we aim at stating that the commons is the best solution to data governance in the 4IR.

[5] Data commons means not only open data but also collective practices such as the co-creation of data protection solutions to rebalance power between data subjects and data controllers. See Janis Wong, Tristan Henderson and Kirstie Ball, ‘Data Protection for the Common Good: Developing a Framework for a Data Protection-Focused Data Commons’ (2022) 4 Data & Policy 1.

[6] Directive (EU) 2019/1024 of the European Parliament and of the Council of 20 June 2019 on open data and the re-use of public sector information [2019] OJ L 172/56.

[7] Proposal for a Regulation of the European Parliament and of the Council on European data governance (Data Governance Act) COM/2020/767 final.

[8] Open Data Directive, recital 8.

[9] Data Governance Act, Explanatory Memorandum, para 3. See also recital 35.

[10] AI Act, art 64.

[11] AI Act, preamble, para 3.5.

[12] See e.g. Piotr Mikuli and Grzegorz Kuca (eds), Accountability and the Law: Rights, Authority and Transparency of Public Power (Routledge 2021).

Should teachers use sex as a remedy to unfair content moderation practices?

Today (Wednesday 6th July 2022), I’ve presented a provoking (not sure whether thought-provoking) paper called “Should teachers use sex as a remedy to unfair content moderation practices?” at the most fun conference of all i.e. Gikii, in its Icelandic edition.

In May 2022, OnlyFans creator Kitty Lixo revealed how she dealt with the allegedly unjustified deletion of her Instagram account. She tracked down some Facebook employees, slept with them, and her account was promptly reinstated.

A current trend in internet regulation is delegate to private companies the decisions as to who and what is acceptable online. The Online Safety Bill, the Digital Services Act (DSA), and the Copyright in the Digital Single Market (CDSM) are only the best known examples of this.

This paper presents empirical data about content moderation in distance learning platforms and asks: what happens if a teacher spends months to put together entire courses and Zoom, Teams, etc. decide to take down the content or even disable the account? Is having sex with the relevant platform’s employees the only remedy we have?

My suggestions are:

(1) a horizontal right to genuine human review (building on DSA, CDSM, but also GDPR Art. 22, the debate about the human in the loop in the field of lethal autonomous weapons, and the mostly overlooked recent proposal for a Distance Financial Contracts Directive);

(2) a killswitch obligation for platforms/providers of automated moderation tech to disable the tech and allow humans to prevail over technological management.

You can watch the recording of the presentation here


Fourth Industrial Revolution and European Intellectual Property Law

Tomorrow I will present “Fourth Industrial Revolution and European Intellectual Property Law” at the international conference Intellectual Rights: Challenges of the 21st Century, Tomsk State University 9-13 November 2021.

Information on the conference can be found here. The event will be streamed live both in English and in Russian here.

My talk will deal with:

•Fourth Industrial Revolution (4IR)

•IP issues in the 4IR

•Data ownership v sustainability

•Focus on sui generis right AKA database right

•Can we regulation of database rights help open up data thus achieving the sustainable development goals?

•Interpretation of current rules

•Policy considerations (‘de lege ferenda’)

Are We Owned? Abstracts and bios of a conversation on Intellectual Property in the algorithmic society

On Friday 8th October 2021, the University of Stirling will host «Are We Owned? A Multidisciplinary and Comparative Conversation on Intellectual Property in the Algorithmic Society», a conference funded by the Modern Law Review and organised by me with support by Arletta Gorecka (Strathclyde / Stirling).

Here’s the extended programme, including abstracts and bios. At the top you can find the presentations that will be delivered live on the day, following right after there is a group of pre-recorded asynchronous presentations, finally the bios. You can click on the link to watch the videos of the presentations 🙂

9:30 Welcome by Alison Green (Institutional Dean for Teaching Quality Enhancement, University of Stirling)

9:35 Introduction by Guido Noto La Diega (Scientific Coordinator)

Analytical framework: IP in the algorithmic society

Chair: William Webster (University of Stirling)

9:40 Joshua Fairfield (Washington and Lee University), From Owned to Owner: Changing the Core Narratives of Digital Property

How we talk about ownership matters as much or more than the technological affordances of a particular technology. Our narrative imagination around ownership sets the stage, and technology carries it out. What matters is what we want. Consider the example of emerging markets in unique digital property. The underlying technology for NFTs are a loose constellation of cryptographic hashes, a decentralized database for recording interests, and often some form of intellectual property rights. Human desire to own transforms this via a kind of social alchemy into a digital asset that one person is proud to buy, display, invest in, and transfer to others. 

We have had loose constellations of such rights and technologies for years; as noted in Owned, these interests have been drowned by the drumbeat of intellectual property rightsholders concerned about rampant copyright infringement. What makes NFTs different is that they have triggered large enough narratives to force law to take notice and to validate true rivalrous digital ownership. The amount of money involved in NFT purchases, as well as the presence of an element of the asset — the token — that clearly falls outside of intellectual property rights, will create a baseline example of online ownership around which law will coalesce. Currently, law does not support the expectations of owners who invest in these loosely stapled assemblages of technology and intellectual property. But law proceeds by anecdote and example. Until now, domain names have been the best example of pure digital assets, free of the overarching control of IP rightsholders. The example of NFTs will do three things: it will force courts to provide legal backing to ownership rights in purely digital assets, a kind of legal right that has until now been eclipsed by streaming and license models under copyright law. It will push the development of these personal property ownership rights in the face of intellectual property concerns. It will simply not work for the owner of a $69 million piece of digital artwork to be constrained in the same way that one is constrained to own no rights in an e-book or MP3. Once legal protection for digitally unique assets has been established and the hegemony of IP rightsholders broken, these new forms of digital ownership will serve as an anchoring example for other forms of online value, spreading from high-dollar assets to those held by everyday consumers. 

10:00 Marco Ricolfi (University of Turin), The Future of Intellectual Property in the Algorithmic Society

The paper presentation deals with the shift from Intellectual Property to de facto powers. IPRs played an important role for fostering innovation and creativity as the exclusivity on which it was based assumed a balancing exercise by legislature and the mediating role of courts. IoT, Big Data, Machine Learning and AI are based on self-enforcing features which shortcut legislatures and courts. We are still in search for countervailing power which keep in check holders of de facto powers.

10:20 Cristiana Sappa (IÉSEG), Who owns cultural heritage in the algorithmic society?

Cultural heritage institutes (CHI) host collections bearing a cultural interest. The latter makes it essential to ensure the access to them or, at least the communication of information on cultural heritage to a large number of people. This is why, in some way, the community should “own” CHI collections and their pieces. This approach is aligned with the notion of knowledge commons that applies to CHI and that in principle favours access and some uses of the collected material. In a nutshell, this is a traditional approach to the issue, which focuses on access to premises – and therefore, to works -. Such approach did fit with the analogic world and the early phase of the so called communication society.

The said collection are also of interests for all re-users that want to take cultural, educational, as well as economic activities. As an example, market operators dealing with Augmented and Virtual Reality are more than ever looking at this sector. Unfortunately, legal rules do not favour such re-uses and thus hinder the flourishing of commercial initiatives (at least). The practice shows when and how some market players intervene into the scenes and create additional bottlenecks to the viability of the said initiatives. Overall, the community does not seem at all to own cultural heritage for wide exploitation purposes, in particular in the digital realm.

Technology and some social behaviours are always one step ahead to legal rules and approaches of some of the cultural heritage institutes managers. This analysis studies the notion of ownership as applied to CHI collections in a copyright, PSI and cultural goods rules, as well as in a contractual perspective. More precisely it examines how the current algorithmic society focuses on re-uses via the available technology, such as AI, while CHI are still reasoning in terms regularly used in a conventional analogical realm or in the best scenario – under an information society paradigm.

10:40 Q&A

11:00 Coffee break

Owning artificial creativity

Chair: Richard Haynes (University of Stirling)

11:15 Ignas Kalpokas & Julija Kalpokiene (Vytautas Magnus University), Beyond Anthropocentric Creativity: AI, Humans, and Going beyond Copyright

The aim of this paper is to de-anthropomorphise the notion of creativity and, consequently, to re-conceptualise the notion of copyright for an era of creative AI. It will be suggested that for the era of post-human creativity, post-copyright regime is required. While copyright is suitable to regulate human authors, when we talk about autonomous artificial intelligence, an alternative to copyright should be sought. The paper will open with a foray into post humanist theory in order to critique anthropocentric approaches to creativity. Notions of human autonomy and exceptionality will be challenged on normative grounds and in terms of practical impossibility, particularly vis-à-vis practices of datafication and the ascent of AI (drawing on e.g. Gunkel 2018; Braidotti 2019; Lupton 2020). Support will also be drawn from neuroscience-informed approaches that disprove any ideas of depth in human conscience and objective cognition of the environment on which human exceptionality, including in creativity, could be founded (see e.g. Charter 2019;Hoffman 2019;Oliver 2020). As a second step of the argument, potential for AI creativity will be explored, focusing on technological capacity (as informed by e.g. Du Sautoy 2019; Miller 2020) and the possibility of a non-anthropocentric framework for establishing creativity, thereby setting up the scene for the third, and most important, consideration: what, if any, results of creativity are to be protected and, if so, how. In order to assess the suitability of copyright for regulation of post-human creativity, philosophical underpinnings of copyright will be considered. Then, it will be argued that once not only human authors are competing in the market, copyright is not capable of effective regulation as the incentives offered by it are not relevant to post-human authors. As such, it will be proposed that for the era of post-human creativity a new regulatory regime is preferable.

11:35 Burkhard Schafer (University of Edinburgh) & Erica Fraser (Canadian Intellectual Property Office), Computational Creativity is AI having fun: the law of explainable and creative Ais

The talk aims to bring together two issues in the law of AI that have each attracted some attention, but which so far have not been brought into an exchange with each other, explainable AI and computer creativity. They seem to  address different research communities, and to raise very different legal and ethical issues. Indeed, they seem almost diametrically opposed: Creativity is often understood as an ephemeral quality that eschews explanation, it is what happens when rules and predictability run out, or as Einstein put it , “when intelligence is having fun”. The type of decision that we are concerned about when talking about explainable AI also seem to be not much fun: A rejection of a credit application, the classification of a human as “untrustworthy” etc. These are very different from the “decision” to use a colour scheme to capture an emotion or a gestatswitch that enabled a new invention. The research on explainable and interpretable AI to prevent unfair or discriminatory decisions has focussed exclusively on the former. In the field of creative AI by contrast, the legal issue at the forefront of the discussion has been the status of the results of computer creativity, in particular if they are patentable. Questions surrounding the patentability of computer generated innovations are now discussed in the literature, albeit yet to be definitively resolved. No attention has been given however to the question if for these legal purposes, creative AIs also require (a form of) explainability of their processes and procedures, and if so what type of explanation. We will show how in traditional patent validity or infringement trials, inventor testimony, an inventor explaining their creative process, can be probative with respect to issues such as defining the common general knowledge, obviousness, utility and sufficiency of disclosure. Where AIs or robots take an increasing role in the inventive process, this then leads to the question if they too should, and could, be interrogated in the same way and provide probative value, “AIventor” testimony. What type of “explanation” of their creative processes is appropriate for legal purposes, how do they differ from the better studied examples of explainable classification decisions, and how could that type of explanation be achieved?

11:55 Q&A

12:15 Lunch

Critical perspectives

Chair: Rowan Cruft (University of Stirling)

1:30 Matthew David (Durham University), Living in a vacuum –9000 patents or £9000 tuition fees?

In 2017 Dyson Appliances announced its intention to revolutionise UK higher education with a new model of practical learning, no fees, paid internship/apprenticeships, and the chance to develop patentable (by the company) applications of ideas whilst gaining a degree. The company’s 9,000 patents (far greater than the number filed by any university) are cited as evidence that original thinking is better found in that company than in any established academy. These 9,000 patents are then set against the £9,000 plus fees charged by established English universities. Ownership over innovation, as secured by IPRs, allows the Dyson Institute to pay students. Is such control over innovation a suitable foundation for higher education? Can principles of universal knowledge, and the method of scientific innovation by which ideas are made available for testing and challenge, be sustained in conditions of total IP control? Might is not always right, but it may not always be wrong either!

1:50 Aleksandra Nowicka (Westminster Law School), The issue of the artificial intelligence being based on the algorithmic bias – the analysis through the feminist lenses’

We live in a men’s world. The world created by men for men. Even though women amount to the majority of world’s population, they hold only a small percentage of senior chairs. The academic research is dominantly conducted through the male perspective. The research data is analysed through the lens of its potential effects on men, without introducing a female element. For example, it was documented by the Centre for Gender Medicine at the Karolinska Institutet in Stockholm, Sweden, that cardiovascular diseases have been studied exclusively with male patients as research models, whereas, they develop differently in women than in men.

It is still presented that there still exist a hegemonic intersectionality between the categories of ‘man’ and ‘human being’, and that it is a prevailing starting point for many traditional researches (Hanmer 1990). It also takes place when we use search engines. They were programmed in a way to put an equal sign between ‘human’ and ‘man’.

Do we support our research via using search engines such as Google Scholar or Researcher? Yes, we do. It makes our lives easier, but does it give us a full picture and complex research results? No, it does not. Why? Because those engines use artificial intelligence (AI) based on algorithms. The algorithms which were again created by men for men. Therefore, we can observe so-called algorithmic bias.

It shows the necessity of the “revolution” in the tech world of AI by letting more women to be the coders. Women should be included in AI creation and programming. Their perspective will allow to minimise or even exclude algorithmic bias, also via stopping a traditional merging of ‘man’ and ‘human being’.

2:10 Q&A

2:30 Coffee break

Will technology save us from technology?

Chair: Mo Egan (University of Stirling)

2:55 Kwaku Adu-Amankwa (Strathclyde University) & Angela Daly (University of Dundee), Whose Digital Property Is It Anyway? An Interdisciplinary Perspective on Intellectual Property, Technical Protection Measures and Digital Manufacturing Supply Chains

Digital supply chains provide the necessary virtual materials and content for the production of objects in Industry 4.0 smart manufacturing contexts, bringing important advantages over traditional physical supply chains in terms of efficiency, efficacy and sustainability. A digital file comprising an object to be manufactured can be sent along a digital supply chain almost instantaneously to a smart manufacturing site which can then construct the object on-demand using e.g. a 3D printer. However, digital supply chains pose new risks as the data which travel along such supply chains may both comprise IP-protected material and challenge the protection of the IP of others. There is also the risk of data breaches which may involve the disclosure of IP-protected information.

One way of addressing IP security issues in digital supply chains is using technical protection measures (TPMs) within them. However, as with TPMs introduced to protect copyrighted content in digital scenarios from the 1990s onwards, TPMs in digital supply chains may over-protect IP contained within them and not pay enough regard to IP exceptions. Furthermore, in digital supply chains, a range of IP rights may be at stake, including copyright but also patents, design rights and trade secrets, which add complexity to the operation of TPMs inasmuch as different IP rights have different subject-matter, scope and exceptions which may be difficult to address digitally.

Accordingly, this presentation seeks to critically assess these complexities when applying TPMs to IP (especially non-registrable rights such as copyright and trade secrets) within digitally-enabled manufacturing supply chains from an interdisciplinary Law and Engineering perspective. We will give an overview of what digital supply chains are, their benefits and risks, before turning to the kinds of IP security scenarios which play out in digital supply chains. We will look at the use of TPMs by different parties in the digital supply chains as IP security measures, and increasingly the use of automated IP security measures using algorithmic methods. We will discuss whether they strike the right balance between the competing interests of different digital supply chain actors or whether they overprotect the IP at stake and therefore are another form of digital enclosure, diminishing the democratizing and sustainability gains promised by digital supply chains and smart manufacturing in Industry 4.0.

3:15 Janice Denoncourt (Nottingham Law School), Corporate Intellectual Property and Technology in an Algorithmic Society

The research in this paper considers the perspective of the corporate IPR owners and elaborates how company law and director accountability could better address the issues facing digital tenants, those who no longer own or control an object or personal data.  

            Within the digital tenant debate, the theory of corporate purpose and law regulating directors’ duties give rise to tensions between the disciplines that need to be reconciled.  From a corporate law perspective, company directors are mandated by law to promote the success of the company according to section 172 Companies Act 2006.  The UK Act does not define ‘success’ nor include any reference to ‘technology’ or ‘intellectual property’.   A company director must act in a way s/he considers, in good faith would be most likely to ‘promote the success of the company for the members as a whole’.  However, in doing so directors must have regard to a variety of matters set out in the CA 2006 including ‘the community’, among others.

            Thus, adopting an interdisciplinary law methodology, my research explores potential reforms to company law in the context of IPRs and technology to improve corporate accountability in an algorithmic society.    While it is generally agreed that ‘ethical, safe and inclusive AI’ should guide corporate behaviour, how that is activated in practice by firms is problematic.   There are potential social, safety, inclusivity as well as economic consequences.  The emergence of digital tenants is perhaps an example of an unintended consequence.  I study the USPTO’s Public Views on Artificial Intelligence and Intellectual Property Policy (October 2020) and the Montreal Declaration of Responsible AI (December 2019) and relevant literature with a view to determining relevance and usefulness in a corporate governance context.     This paper will update traditional corporate governance theory on corporate purpose and directors’ duties to better address contemporary practice in an algorithmic society.  The conclusions and recommendations will assist corporate regulators and company directors develop a more nuanced understanding of their mandated duty to promote the success of the company, balanced against what this means for ‘the community’ (s 172(1)(d) CA 2006) in a contemporary algorithmic society. 

3:35 Q&A

Solutions

Chair: Guido Noto La Diega (University of Stirling)

3:55 Maria Lillà Montagnani (Bocconi University) & Alina Trapova (University of Nottingham), The GDPR to the rescue of algorithmic copyright enforcement

The newest EU copyright directive (Directive2019/790) does not explicitly mention algorithmic enforcement. Nonetheless, the only way for intermediaries to effectively fulfil with their obligations under Article 17 is to install algorithmic means of content moderation. The academic literature has been vocal about the need to safeguard user rights–platforms are likely to resort to ex ante filtering to comply with Article 17, thereby risking taking down completely legitimate content, such as works falling within the copyright exceptions and limitations. Hence, intermediaries, legally obliged by EU legislation, are geared with overprotective algorithmic enforcement tools, like YouTube’s ContentID. They “delegate” the control of what content stays and goes down to algorithms and implement the legislator’s request through algorithm-decision making processes. Although Article 17 stresses the importance of maintaining the availability of works uploaded by users not infringing copyright, this seems to be a mere wishful thinking hinging on the hope that intermediaries’ algorithms are extremely good at making sensitive value judgments on copyright law. This scenario begs two inevitable questions as to who the real owner of users’ expression is and, once the intermediaries’ control over user rights is ascertained, what tools for re-appropriation are left in the users’ hands. In our analysis, we approach the problem by drawing on Article 22 of the GDPR (Regulation 2016/679)–the data subject’s right to object to algorithmic decision-making. Our research design infuses this unambiguous obligation into copyright law by demonstrating that a parallel is possible since digital copyright content and personal data are two different layers of information; thus, subjecting the copyright discussion to the GDPR paradigms in this respect is reasonable. This research has broader implications –what is at stake with algorithmic copyright enforcement are fundamental rights, the curtailing of which should not be a decision left in the “hands” of an algorithm.

4:15 Sean Thomas (York Law School), Embedded software and sale of goods: a suggested legislative framework for English and Welsh law

IP law (and its cognates, relating to data) tends to dominate any discussion concerning software. Given that software is embedded into many tangible things, such an approach is potentially problematic. Can the concept of a sale of goods resist the colonialism of IP law? If it can (and it is suggested that it both can and should), one may further wonder whether sale of goods law can provide a functioning regime for transactions involving goods with embedded software. The meaning of software vis-à-vis goods has long been uncertain: and it is unclear whether Computer Associates UK Ltd v The Software Incubator Ltd,[1]will clarify the issue. Substantial problems exist when the implications of the Sale of Goods Act 1979 provisions concerning the obligations to pass good title and the protection of quiet possession are recognised.[2] The rise of product-as-service and moves to remove ownership as a focal point of a goods transaction,[3] arguably requires a new legislative framework for commercial transactions of goods with embedded software. Such a call is further supported by the removal of consumer transactions (by virtue of the Consumer Rights Act 2015) from the reach of the SGA, the generation of a new sui generis form of sale,[4] as well as doubts about the value of even a skeleton SGA.[5]

This paper considers what a commercial sales law for goods with embedded software should look like. A critical examination of EU Directives 2019/770 and 2019/771 is also provided, with the suggestion that that approach is inappropriate. This paper will propose an alternative legislative framework that fits within the general tenor of English and Welsh commercial law, without being beholden to treating the issue solely or even primarily as a problem for IP law.

4:35 Q&A and conclusions

Asynchronous presentations

AI AND COPYRIGHT

Ying Ye (Durham University), Computer and Human Being: Who Is the Author Of AI-generated Content?

Topics of AI-generated content focus on the ownership under the current copyright system. However, not enough attention is paid to the actual generation process of AI-generated content. The current copyright law is built on the assumption that human beings are the subject of creating a work. And the authorship is the important coordinate point for locating the ownership of copyright. Generally, the act of creating a work allows a person to directly obtain the authorship. Based on the legal regulations or contractual assignments, a person can also gain copyright. These two roads are the general approaches for original acquisition. When the rules applied to AI, it may cause confusions or even not function. When the contributions made by humans and AI in the content generation process are indistinguishable, or human input is apparently “minimal”, how to confirm the authorship becomes a problem. The ambiguity of the authorship also makes the ownership confusing. This paper argues that the key to identifying authorship is to clarify the generation process of AI-generated content. It refers to two typical cases of AI-generated works and copyright issues in Chinese courts to illustrate how human teams and AI work together to generate new content. It provides solutions to identify the authorship under the framework of Chinese copyright law, which also gives a possible approach for the international community to address challenges of AI creations and related copyright issue.

Vincenzo Iaia (Luiss University), To be, or not to be…original, that is (one of) the main questions concerning AI-generated works (discussant: Professor Burkhard Schafer)

“To be, or not to be original, that is the question”, this would have been one of the main dilemmas to which Hamlet would have addressed his attention if he were a copyright scholar during the age of Artificial Intelligence (AI). The originality assessment would have been focused on the works generated by AI, especially writings, music and paintings, as confirmed respectively by the press article published on The Guardian drafted by the AI-system named GPT-3, the playlist “AI-composed music” on Spotify and the Rembrandt painting produced 350 years after his dead. Such list is destined to widen together with the increase in computational capacity of AI systems. Of course, the here mentioned question was not treated by the Shakespearean character for well-known reasons, but many contemporary and authoritative scholars have extensively discussed this intriguing topic. More precisely, it has raised almost three intertwined questions: (i) whether AI-generated works deserve copyright protection, (ii) in case of positive answer, which natural/legal person should be considered as the author, (iii) how the “creative capacity” of AI can influence the originality standard and, more broadly, the creative market. This paper will try to provide some answers by carrying out an analysis of the antagonistic interests at stake. From one side, high-tech companies claim to be rewarded for their investments, invoking the copyright incentive theory and the principle of technological neutrality. From the other side, it should be considered the opposite interest of the creative market to preserve its competitiveness, as it could be menaced by the proliferation of monopoly rights on the ever-increasing AI-generated outputs. In the attempt of striking a sustainable balance, the opportunity of a further expansion of copyright on AI-generated works will be evaluated, together with the assessment of some existing alternative legal tools, like unfair competition, patents and contracts.

Alkistis Kympari (University of Nantes), Recognizing Copyright Rights for Artificial Intelligence (AI): A Discussion on the Challenges of Ownership and Access in the Algorithmic Society from a Policy and Legal Perspective (discussant: Professor Daniela Caterino)

While AI technologies grow more and more proficient, the issue of AI creativity has become a challenging question among different disciplines. As famous examples of AI projects manifest, AI-generated creations have the potential to be put alongside works of human ingenuity, prompting debate on whether these creations can be classified as copyright materials. From a legal perspective, the concepts of authorship and originality are both inherently linked to humans marking their own spirits upon a work. That being said, the copyright doctrines are regularly revisited in order to adjust to new technological realities and several ideas on how to protect AI works are explored; should self-learning robots be granted legal personhood, AI could be considered to be itself owner of copyright. In this case, unless a well-defined legal scheme can be found, there is a risk that copyright protection could produce more uncertainties than it could solve. The circulation of such “AI-authored” works, whose parentage cannot be attributed to natural persons, could easily cause further confusion among users that are already perplexed about how to utilize protected material in the digital era. In other words, this approach could result to a new generation of orphan works, the latter often being considered at the intersection of copyright and public domain. While the innovative  potential of machine learning algorithms processing large amounts of data is indeed hindered by complex copyright issues, it remains debatable whether AI protection constitutes a justified and beneficial course: would we be creating more barriers, when we could instead engage with policies that regulate and promote openness and data sharing? Understanding the policy interests that motivate AI protection could assist the recognition of existing legal challenges and reveal appropriate solutions. What is the legal necessity for protecting AI works and which legal tools could best serve this need?

Israel Cedillo Lazcano (Universidad de las Américas Puebla), Analysing the Creative Potential of Artificial Intelligence through its “Moral Control” of the Market

The rise of Artificial Intelligence (AI) within the so-called Fourth Industrial Revolution (FIR) and its creative potential have been discussed in different academic sources and forums throughout the world. For example, institutions, such as, the World Intellectual Property Organization (WIPO) have organized different consultations and forums to discuss the legal nature of the works developed by AIs that go beyond simple expert programmes. Most of these efforts tend to rely on accurate analyses that emerge from questions like “who controls the economic rights?” However, in rare occasions, we stop to consider who has the right to make modifications on a work created by an AI or who decides that this latter is ready to be displayed and/or offered within the market just as we did with the case of Banksy’s “Girl with Balloon.” Of course, to answer these questions, we have to go beyond economic rights and analyse those moral rights that could be involved in the display and the offer of the work beyond the simple paternity right. With these elements in mind, the present paper will offer an analysis that will complement existent works highlighting the relevance of moral rights in the control over the work within the market to determine: 1) who can be labelled as the “Kantian master mind” behind the work and 2) what do we need to consider an AI as a potential author.

DATA GOVERNANCE

Simon Geiregat (Ghent University), Data Ownership: the EU is on it (and it’s not IP!)

Since long, ownership in data is a touchy subject in European doctrine. Although property rights in software and information were already extensively discussed at the end of last Century, the debate only took its current shape after calls of German stakeholders who favoured the idea of exclusive rights inindustrial data – increasingly more valuable “assets” in our present-day digital economies. Courts in Western Europe seemed reluctant to acknowledge ownership in data. Moreover, it is believed that legal exclusivity in data could not be sufficiently guaranteed by existing mechanisms, like contract, patent, database and trade secret law. Consequently, it was debated whether exclusive rights in datashould indeed be created, and whether they should classify as either property or IP rights. Quite soon after the discussion had emerged though, there was broad consensus that neither data ownership nor data producer’s rights correspond to an actual socio-economic need, and that a more targeted creation of data access rights might instead be the way forward. Although it has not explicitly ruled out the option of creating ownership-like rights, the European Commission seems to agree that introducing data access rights would be beneficial to the “Data Economy”, notably in order to strengthen the position of consumers as well as SMEs that require access to certain data to enter a market1. Normative interventions, like the 2021 Data Act proposal, are therefore expected to enhance data sharing and data portability. It looks like the creation of concurring access and usage rights have replaced the vain search to attribute exclusivity to data, which are non-rival assets by their very nature. Combining all this with the portability right in the 2016 GDPR2 and the digital restitution rights in the 2019 Digital Content Directive3, it furthermore seems, with hindsight, that the European legislator has already started the enactment of the rights formerly known as data ownership…

Giulia Schneider (Sant’Anna School of Advanced Studies),Licensing Personal Data in Europe: How Data Protection Can Re-Match Intellectual Property Law to Innovation Objectives in Data Markets (discussant: Professor Cristiana Sappa)

Data sharing has lately become a key concern with the objective of boosting the data availability and related innovation results within the European digital single market. The importance of access to data has been lately highlighted by the European Commission in its “European strategy for data”. Here, the creation of outright “European data spaces” concerning the most important sectors of the economy are being incentivised. Under the new strategy for data, data pools shall be as “open as possible” and as “closed as necessary”, so as to promote data re-usability and analysis across different sectors of the economy.

Against this backdrop, the paper enquires the limits and the opportunities provided by the European data protection and intellectual property frameworks in respect to the set policy objectives of stirring research over large datasets and research platforms’ interoperability. It ultimately demonstrates the possible relevance of data protection law in the re-alignment of intellectual property law with innovation objectives in data markets.

After having briefly identified the main provisions under data protection law and identified the intellectual property rights which are likely to hamper data pooling practices, the paper enquires a further barrier to data sharing given by the lack of interoperability standards of shared personal data. It thus enquires the opportunities of the establishment of licensing schemes regarding personal data, questioning which type of licenses (e.g. FLOSS; license granted also for commercial or only non-commercial uses) would be desirable to use for data-driven innovation purposes. 

Drawing on these findings, the third section will illustrate the possible synergies between data protection law principles and traditional intellectual property schemes for the achievement of sharing and interoperability objectives. In particular combined schemes directly resulting from traditional intellectual property schemes and data protection principles, as the principles of purpose limitation, fairness and transparency, will be explored. In this perspective, the study will provide a reconsideration of i) the FRAND terms, ii) grant back clauses and iii) technical standards of openness regarding the sharing of research valuable personal information, , vis à vis data protection principles, employed as a design benchmark for licensing schemes regarding data.

The study ultimately demonstrates the possible relevance of data protection law in the re-alignment.

Tommaso Fia (European University Institute), Resisting intellectual property rights in data. From ‘data governance’ to ‘data processing governance (discussant: Associate Professor Janice Denoncourt)

Data is not a widely available resource. Quite on the contrary, it is subject to the ability of several actors to control it and prevent others from doing so(‘data ownership’). In algorithmic societies, however, there is urgent need to open up access to data and encourage data sharing. My paper examines a self-contained aspect of the data ownership matter, that is intellectual property rights (IPRs) in data, and explores alternative modes of data governance. It comes in three consequential parts. First, it concisely examines how intellectual property rights (IPRs) stretch to data –particularly large-scale datasets. Exploring IPRs in data has become a long story in the legal literature in the last five years. Thus, the paper succinctly reviews the literature that has interrogated how two IP standards (ie database rights and trade secrets) apply to data. Second, the analysis moves beyond a purely descriptive approach to consider how IP laws can be restricted in light of the need to open up data access. Instead of adopting a rule-based approach, it singles out a set of countervailing principles that can help restrict(rather than rejig) IP(over-)expansion to data. A novel interpretation of the numerus clausus principle and the social function of (intellectual) property can help the legislatures and courts in this respect. Lastly, and more interestingly, I identify the tools for building alternative modes of data governance that apply horizontally to various sectors. In doing so, I maintain that the interpretation of the commons as a humanitarian-oriented force can prove useful as to decentralise control over data and foster data access. Drawing inspiration from key concepts within data protection laws, I argue that ‘data processing’ (rather than just ‘data’) should be understood as an infrastructure(rather than a commodity) that can be accessed for a legal basis and a purpose.

Jozefien Vanherpe (KU Leuven), Know your audience: securing transparent access to audience data in the digitised music industry (discussant: Professor Angela Daly)

Over 400 million people currently have a paid subscription to a premium music streaming service, averaging a total number of yearly streams surpassing a trillion. In 2020, a single song garnered 1.6 billion streams on Spotify alone. The most successful artist has collected over 36 billion streams so far. The amount of revenue amassed by music streaming on a global level is simply staggering.

Valid questions arise as to how these data are managed in practice as well as whether composing and/or performing musicians may access relevant data to ensure correct and fair payments from their corporate partners (primarily music publishers and record companies). While accounting obligations in the context of copyright exploitation contracts were partially harmonised through Article 19 Digital Single Market (DSM) Directive, significant obstacles remain in practice. Secondary contractual relationships between musicians’ corporate partners and streaming services are often rather opaque at best, leaving musicians in the dark as to whether they receive fair remuneration for their artistic contributions. A linked issue pertains to the data that allow direct contact with listeners, such as a mailing list compiled over the years, or a band website filled with relevant content. Since such data are usually managed by corporate partners, save in a true do-it-yourself (DIY) context, musicians who wish to seek new horizons may be constrained by the prospect of losing direct access to their fanbase. Thus, exclusive access to data risks functioning as a switching cost that risks locking in musicians, leaving the promise of disintermediation through digitisation to sound rather hollow.

Taking account of the particular contractual dynamics in the digitised music industry, this presentation tackles the question whether and how the legal framework may contribute to securing access to relevant audience data for musicians, both throughout their contract and afterwards.

Jacopo Ciani Sciolla (University of Turin & ESCP Business School), EU Regulation on European data governance and data ownership (discussant: Professor Maria Lillà Montagnani)

Understanding who owns data or whether a hypothetical property right on data should be introduced in the EU legal framework are pivotal issues on the Digital Single Market Strategy for Europe political agenda. Lacking legally specified ownership rights over data, who holds it may effectively use technical and contractual measures in order to prevent others from accessing it. This scenario will shortly be supplemented by the EU Regulation on European data governance, currently at the proposal stage. Although the accompanying report denies any purpose to grantor modify substantive rights over data, actually the proposal seems to introduce a form of control over it. Indeed, the draft provides that public sector bodies, persons admitted to data reuse and intermediary data sharing platforms should legitimately “take all reasonable measures to prevent access to the systems where non-personal data is stored, including encryption of data or corporate policies”, in order to prevent “unlawful access to non-personal data” (Recital 18). The aim of the research proposal under submission is to understand possible implications of this normative change. The provision raises several issues. Which is the scope of the “reasonable measures” provided by the legislator as data control tool. Is it broad enough as to take the form of both legal and technical measures? Could the latter be similar to the technological protection measures shaped by the Infosoc Directive for preventing unauthorized access to copyrighted works? Shall the scope of application of this regime be limited to public sector information or could anticipate the next steps to be brought by the Data Act in relation to non public data? In light of these new regulatory developments, the data ownership issue seems still a long way from being resolved.

A FOCUS ON THE USER

Amy Thomas (University of Glasgow), Are we pwned? (discussant: Dr Gaetano Dimita)

The act of reselling a video game is an inflammatory topic for game creators. In the past, they have described resale as: more problematic than piracy, a ‘critical situation’, something that should be actively fought against and ‘stabbed in the heart’. Legal support for these anti-resale sentiments may also be echoed in copyright law. However, this support is highly variable depending on whether a game is distributed physically or digitally, or whether it’s considered as a good or a service. Any, all, or none of these options may apply as gaming business models rapidly evolve. Given the uncertainty about whether the law will support a game creator’s position on resale, is it possible to construct an idealised user as a ‘non-reseller’ outwith the law?

To answer this question, this research looks beyond a strictly legal perspective of ownership and instead to the role of language in constructing the user through end user licensing agreements (EULAs). Using critical discourse analysis, the research examines how ownership (or pwnership, in this case) is negotiated from a sample of EULAs in this industry.

The research finds that the strong anti-resale sentiments present in the video games industry filter into micro-level linguistic and semiotic features of the EULA (and indeed, vice versa). In comparison to other contentious issues, such as the regulation of user-generated content, discourses surrounding resale are very consistent throughout the industry, regardless of the sociocultural background of the game creator. Whilst restrictions on resale are often implicitly tied with the commercial value of the work, it is also positioned closely with concepts like fraud and piracy, inferring a logical connection between them. In combination, these textual constructions can play a strategic role in legitimating and justifying anti-resale sentiments at industry level.

Giovanni De Gregorio (University of Oxford), João Pedro Quintais (University of Amsterdam, Institute for Information Law) & João Carlos Magalhães (Alexander von Humboldt Institute for Internet and Society & University of Groningen), How Platforms Govern User Copyright-Protected Expression: Are Users Owned? (discussant: Professor Dan L Burk)

Online platforms provide the main point of access to information and other content in the digital age. They foster users’ ability to share ideas and opinions while offering opportunities for cultural and creative industries. In Europe, ownership and use of such expression is partly governed in part by a complex combination of legislation, sectoral self-and co-regulatory norms. To an important degree, it is also governed by private norms. By adopting Terms of Service (TOS) and Community Guidelines, platforms unilaterally set use, moderation and enforcement rules, structures and practices –including through algorithms –that govern how users access and disseminate protected content. This private governance of essential means of access, dissemination and expression to (and through) creative content places users (including user-creators) in a precarious position. In a very real sense, their online expressive activity is “owned” by platforms. As platform power grows, EU law is adjusting by moving towards enhancing the responsibility of platforms for content they host. The primary example of this is Article 17 of the new Copyright Directive (2019/790), which fundamentally changes the regime and liability of “online content-sharing service providers” (OCSSPs). This complex regime, to be complemented by rules in the upcoming Digital Services Act, sets out a new environment for OCSSPs to design and carry out content moderation, as well as to define their contractual relationship with users, including uploaders/creators. The latter relationship is characterized by significant power imbalance in favor of platforms, calling into question whether the law can and should do more to protect users. This paper addresses the governance of copyright-protected content by online platforms and its effects on users’ rights and remedies. We explain how private ordering by platforms enables them to effectively “own” their users’ expression, and what can and should be done to address this.

Georgia Jenkins (Queen Mary, University of London), A secondary market for digital content access rights

It is paradoxically difficult for users to access digital content: online platforms offer instant consumption and yet users ‘own’ nothing.[6] They may demand access at any time or location and in most formats but are subject to restrictive licensing terms. Conversely, copyright holders face a perpetual war against infringement as users distribute and access an unlimited number of perfect copies of digital content more easily. In response, they use technological protection measures (TPMs) to digitally ‘lock’ works.

Automated copyright enforcement spells further disaster. Article 17 Digital Single Market Directive shifts liability to content-sharing service providers. Intermediaries must ensure that works communicated by copyright holders are not uploaded or reuploaded following take down. Despite the ban on general monitoring, the standard of best efforts is ambiguous. Intermediaries can only comply by using artificial intelligence (AI) to decide infringement. However, the notion that AI could determine copyright disputes without the judiciary is chilling, particularly for user rights and the public domain.

This presentation reframes the doctrine of exhaustion as an exemption to rebalance the copyright matrix in the digital era. It reviews EU legislation and deems it lacking considering the internal market context. The presentation proposes that recent CJEU jurisprudence hints at an overarching interpretation of exhaustion as an exemption which could loosen the Gordian knot and help reconcile users and copyright holders in the digital environment.

ALGORITHMS

Daniela Caterino (University of Bari), Technology Protection Measures over Copyrighted Contents and “Intelligent” Algorithmic Control in Remote Teaching Activities

During the Covid-19 pandemic, distance learning activities have registered a unique upward trend, as they represented the only lawful option for ensuring the right to education while preserving the health of all those involved in the learning chain, namely students, teachers and administrative staff. Beyond the heated debate concerning the best balancing solution between these concurring fundamental rights, remote teaching activities necessarily entail two relevant acts: (i) the more or less conscious choice of an online platform through which such activities can be carried out (ii) the selection of the study material(i.e., book chapters, articles, movies, whose copyright has not yet expired) to be uploaded on the chosen platform, in order to be shared with students. These acts are much more intertwined than it may appear, since each platform has its discretionary terms and conditions about the control over the uploaded contents, as well as proper technical and legal tools to escape from the hosting provider liability regime for copyright infringements, as settled by Art. 14 of the InfoSoc Directive and by Art. 17 of the yet-to-be implemented CDSM Directive. However, despite Art. 5.1, let. c, of the InfoSoc Directive and Art. 5 of the CDSM Directive oblige Member States to provide educational exceptions or limitations to copyright, the technical protection measures, even the most sophisticated ones based on machine learning systems, have difficulty in recognizing legitimate uses. Hence, in this paper I will try to assess the mode of exploitation of the in-copyright study material that online platforms can lawfully carry out in compliance with the principle of fairness in the European digital single market. Then, I will evaluate the compatibility of deploying fast technological protection measures over copyrighted contents based on (likely) “intelligent” algorithmic control with the prohibition of a general surveillance obligation; as well with the need of complying with the puzzle of copyright exceptions and limitations to pursue other concurring interests, especially the educational one.

Jacopo Menghini (University of Bologna), The challenge to parody’s detection in algorithmic society: can algorithms cope with the parody exception?

This abstract aims  to  focus  on  the  parody  concept  which  traditionally  represents  a  high expression  of  art,  marked  by  stinging  irony.  The  rudiments  of  parody  in  legal  matter  are  quite uncertain and not precise. The search is divided in five parts: the first one inquires the organization of the matter with specific theories in literary field; the second one shows parody as free expression of  a  comic,  humorous  or  satirical  thought,  remarking  constitutional  protection  and  limits; In  fact, there are legal limits on art. Some forms of artistic expressions, like satire or parody, can have a bad effect  upon  important  constitutional  rights:  the  honour,  the  reputation,  the  personal  identity. Therefore, art is not legibus soluta. The third part introduces the notion of parody like the work of talent  in  the  copyright  field,  the  exposition  faces  the  different  positions  of Italian  doctrine  and  the possibility that parody will be recognized as an independent and autonomous work; the fourth one distinguished parody in according to the functions that it performs, inquiring its different expressions in some juridical cases and doing a comparison between the adopted solutions of Italian courts and American jurisprudence; the fifth part aim is to examinate the parodic matter inside the web, reported at the most recent European directives. From Deckmyn case on(C201/13),the European Directive 2001/29/EC(InfoSoc) needs to deal with an actual problem about legal informatic and the issue of online hate speech: the elaboration of parody could degenerate into hateful messages characterised by a markedly defamatory or discriminatory aspect. Authors could ban a parodic transformation of the work  if  the  parody  has the  only  purpose  to defame  or  discriminate. Parodic true  underlying meaning may be easy for humans to detect, but it can be very challenging create a system of detection in the algorithmic society.

Sevra G. Guzel (Hertfordshire Law School), Unwanted Judges of Online Legality: The Dark Side of Automated Decision Making (discussant: Professor Marco Ricolfi)

This paper conducts an in-depth analysis of primary sources such as EU legislation and CJEU and ECHR case law, and recent secondary sources to critically assess the crucial problems with algorithms as decision-making authorities. Online platforms are one of the most important intermediaries that automated and algorithm-depended decision-making takes place. In the wake of the COVID-19 pandemic, these platforms became the primary source for users to exercise their right to freedom of expression and the right to receive and impart information and ideas that is enshrined in the Charter of Fundamental Rights of the European Union (Charter). However, upload filters placed in these online platforms are enforcing IP rights at the expense of users’ fundamental rights. These context-blind filters ignore the lawful uses of copyrighted work and automatically block or remove all content that includes a technical infringement and filter out legal communications covered by the Charter. On the darker side, AI algorithms of these filters can learn to treat unfairly to the groups of online users if their views were poorly or misrepresented in AI training data and filter out expressions of members of racial and linguistic minority groups. All this decision-making is taking place within a black-box as these filters are unable to give reasoning regarding the enforcement decisions about the content they examined. Therefore, there is a great risk of upload filters and online platforms to become judges of online legality as they decide the fate of the content.

To prevent this outcome, this paper suggests an ideal implementation of filtering systems that discards the adverse effects of their decision-making to ensure that the upload filters can be implemented in a way that is compatible with the fundamental rights. Moreover, it contributes to the discussions of the Digital Services Act by extending the knowledge regarding the problems with online platforms.

Alessia Palladino (Suor Orsola Benincasa University of Naples) & Massimo Farina (University of Cagliari), Intellectual property rights in the era of “artificial” public decisions: towards the effectiveness of the right to explanation

The relationship between artificial intelligence and public decisions has been increasingly questioned throughout the last decades by legal scholars, to prevent any unfair compromise of fundamental rights. However, considering that public administration may not be able to develop the necessary technologies in house and will have to procure them on the market, several concerns about intellectual property rights consequently rise. On the one hand, some argue these rights may stimulate an unprecedented dominance of private operators upon the whole public decision-making process. Others, on the other hand, claim the overall disclosure of the source code, as well as every trade secret involved, as the most appropriate counterbalance to ensure higher levels of transparency, freedom of information together with civic engagement, in coherence with Open Government policies. This work aims to investigate the debated morphology of intellectual property in the age of algorithmic administration. By framing the existing rules and the relevant literature, the dimension of the “right to explanation” will be explored: it will constitute the bedrock to verify whether transparency obligations can undermine intellectual property priorities, as the instrumental counterpart of public interest in private intellectual property rights.

Argyri Panezi (IE University & Stanford Law School), Do we own our critical digital infrastructure?

Critical infrastructure is commonly defined as infrastructure the destruction of which has a significant impact core State functions and on society. In the EU context, Council Directive 2008/114/EC defines critical infrastructure as an asset, system or part thereof which is essential for the maintenance of vital societal functions, health, safety, security, economic or social well-being of people. Do we jointly own also the digital components of our critical public infrastructure? Do the digital systems that support our critical infrastructure rely (mostly)on open source or on proprietary software, and why does it matter? The FOSS/proprietary software divide has had both legal and broader implications. This paper focuses on this divide in the context of our (common?)critical digital infrastructure. The paper will explore the following questions: Who owns the software that support the digital components of our critical systems, such our systems of public healthcare, transportation, and also our judicial systems? Does the open-source or proprietary status of software affect the nature of our infrastructures as public or commons? Is there a trade off between openness and(cyber)security? In this paper, I will map the various approaches to criticality of software primarily from the perspective of the State, using the example of the EU and its ongoing efforts to “enhance the resilience of critical entities providing essential services in the EU” (Seethe 2020 Commission proposal for a directive on the resilience of critical entities). I will compare and contrast the State’ s perspective, first, to the perspectives of citizens and civil society, who benefit from but also depend upon critical systems that available as commons, and second, to the perspective of market stakeholders.

PRIVATE ORDERING

Lucio Casalini (Sapienza University of Rome), Commons in the prism of Intellectual Property. Evolutionary profiles of contractual models and circulation of knowledge on the net

Taking inspiration from George Bernard Shaw famous reflection, according to whom “if you have an apple and I have an apple and we exchange these apples, then you and I will still each have one apple; but if you have an idea and I have an idea and we exchange these ideas, then each of us will have two ideas”, my research aims to analyze the category of commons in comparison to Intellectual Property Law (rectius Copyright Law), in the european framework of reference. Commons is elected on a leit motif of research. In particular, the analysis is focused on knowledge as a common immaterial good (id est intellectual and creative works, scientific product et similia), traditionally considered as intellectual property right but radically changing in the context of the new information society, thanks to digitalization. I deep focused potentials and limits of the legal instruments of copyright and, subsequently, the development of new models of contracts on the web. In this perspective, it could ensure the difficult balance between ius excludendi and ius utendi which constitutes the primary objective of a modern intellectual property right. In particular, the focus is on the implementation of Open Access principles, truth e role of the contract: the most flexible instrument for the evolution of the system. Germany is considered the cradle of Open Access European principles: Berlin Declaration is the milestone in the European history of Open Access, dates back to 2003. The purpose of the declaration is to guarantee the maximum possible dissemination of scientific publications through the use of the Internet, which allows the circulation of information, data and scientific products. Considering the oncoming copyright directive transposition, therefore, we should analyze the legal framework of reference and elaborate best practices and new contract models

Johan David Michels (Queen Mary, University of London), Taking back control: Combining IP, consumer protection, and contract law to protect customer data (discussant: Dr Sean Thomas)

Data as such are not property under English law. Indeed, mere information is a poor fit for property rights. Thus, some argue that data exist in a ‘legal vacuum’, where the rules and remedies of property law do not apply (Conway & Grattan, 2017; McGrath, 2016). For example, individuals cannot call upon the tort of conversion to seek return of their data, or to obtain damages when data are lost or corrupted.

Power, like nature, abhors a vacuum. Technology companies like cloud and app providers have stepped in and taken control of customer data. They exercise practical, technological control over the data which customers store in the cloud or process through applications (‘customer data’), as well as legal control, including through standard contracts. For example, cloud contracts often exclude any liability for data loss and may permit the provider to use customer data for their own commercial purposes. As a result, customers are in effect stripped of control over ‘their’ data. (I focus on data submitted by customers, rather than data about individuals. In the latter case, data subject rights under the GDPR balance the interests of individuals, providers, and third parties.)

I argue that English law should recognize the property-like interests that customers have in their data (albeit without explicitly granting property rights). For example, the law could treat cloud services as analogous to an electronic form of bailment, imposing similar duties of care. In addition, recognizing customers’ property-like interests could help challenge unfair contract terms under consumer protection law, as well as unreasonable liability exclusions under the Unfair Contract Terms Act. Doing so would improve legal certainty, support individual autonomy, and help redress the power imbalance between customers and providers in the cloud.

Anthony Rosborough (European University Institute & Dalhousie University), Technologies of Servitude. An Inquiry into Human Agency and Intellectual Property Rights (discussant: Professor Joshua Fairfield)

Ownership is playing a diminished role in our daily lives. Changes in the law, technology and the market have impaired our ability to take ownership in both tangible and digital goods. The service and access model has become the dominant modality through which we interact with technology, the market, and how social norms take shape. This project investigates the role played by intellectual property law in affecting human agency through its contemporary impacts on ownership and personal property rights. It aims to better understand how modern technological design and the legal mechanisms which support it have eroded ownership and dominion over things. The advent of digital copyright law and erosion of the exhaustion principlecomprise the beginnings of this phenomenon. These successes have created the “access right” -a new reality where the philosophical distinction between objects (corpus mechanicum) and the intellectual property which they embody (corpus mysticum) is no longer discernable orrelevant. The product and contract have become fused into one. The elementary claim of this project is that these evolutions in intellectual property laws have allowed rightsowners todesign technologies in ways that rob humans of agency. And thoughintellectual property systems have been deeply implicated in this societal shift away from the agency of personal property, this project looks to their potential role in revitalising it. Itinvestigates the extent to which there are principles of intellectual property which can help us recapture human agency through dominium.

Christof Koolen (KU Leuven), Standard Essential Patents and Access to Technology for Startups: From Patent Hold-up to Technology Diffusion (discussant: Professor Marco Ricolfi)

Consolidating technological knowledge via patent law in the hands of selected intellectual property rights holders has a profound impact on the degree of competition on the market and rivalry among firms. Startups often require access to particular technology stacks in order to participate in high-tech industries. Yet, when essential technology is portioned off by intellectual property rights, entry barriers to the market arise. New market entrants have difficulties competing on a substantive level with patent holders without having access to essential technological knowhow. Standardization processes and Standard Essential Patents (SEP) are geared towards alleviating these concerns by facilitating technology transfers across both horizontal and vertical axes within an industry on Fair, Reasonable and Non-Discriminatory (FRAND) terms. In theory, SEPs should enable small and medium-sized enterprises(SME)to add to innovation. In practice, however, SMEs face significant hurdles when looking to implement standards. First, with limited financial means, startups easily run into problems when navigating and mapping out the applicable patent landscape. Second, young firms encounter difficulties when determining what qualifies as fair and reasonable in the context of FRAND negotiations and hit a dead end when attempting to push for more advantageous licensing conditions. Third, standard implementers are “locked in” once a standard is adopted, thereby bestowing even more market power on the SEP holder while the SME potentially opens itself up to patent ambush tactics. These issues trigger the question to what extent high-tech industries are susceptible to innovation stand stills due to restrictive IPR policies and patent strategies. Against this background, this paper analyses how access to technology –or a lack thereof –influences the dynamics of high-tech markets. By taking an intellectual property law angle towards innovation, the goal is to shed light on how startups experience rivalry within the technology landscape.

NEW CRITICAL APPROACHES

Julio César Padilla Herrera (Universidad de los Andes) & David Felipe Alvarez-Amézquita (University of Tolima),Opennesses and exits for post organic or corporate creators in a post colonised society

Humans seem to be less and less present in the production of works of art. For instance, Fisk, and also others, suggest that corporate authorship tends to dismantle the idea of authorship and introduce the creator as partof the mechanisms of industrial creativity. However, the foundation of copyright is based on human labor, biologically speaking. The idea of an author protected by human rights, though not recent, has been underlined by the ECOSOC since 2005. In the context of AI, Geiger suggests that only humans should be entitled to human rights and therefore copyright protection. Thus, for example, originality is at the core of copyright systems. The author and his or her liberty are fundamental for creativity. These paradigms could collide with a robust group of new post-organic notions, as Sibila and others suggest. An open list of these notions would include the algorithms, the parameterization of both leisure and academic texts, the pre-set sounds for music, the software to improve or create voices, and the standardization of reactions among the audiences that will consume the works. On the other hand, countries like Colombia argue that laws for the promotion of creative industries are the way forward for the development of growing economies. Nevertheless, there is a lack of attention to the dynamics between authors and industries in those laws. This is a problem that, within a framework of creativity based on algorithms, could create the total disappearance of authors. What is proposed here is a study within the theoretical framework of post-organic authorship and corporate authorship of the Colombian law for the promotion of cultural industries. It will be argued that developing countries are opening a new gap for thecultural industries in which authorship rather than fostered could be jeopardized by algorithm-based creativity.

Karen O’Neill (University of Birmingham), AI and copyright: A new materialist perspective (discussant: Associate Professor Matthew David, University of Durham)

AI cannot have copyright over materials that is produced by the phenomena according to s178 of the Copyright, Designs and Patents Act 1988(CDPA).It is contended in this paper that copyright law needs to better understand the meaning of AI before any reforms are made to the CDPA. It is suggested that the current understanding of how copyright matter is formed corresponds with the Newtonian and Cartesian theory on matter. This results in a reductionist approach to copyright works, where works are interpreted as a ‘thing or an object’; evidenced in the very narrow categorisation system within the CDPA. This has the potential for the judiciary to erroneously abstract material as an idea rather than a copyrightable expression, which may have future implications for AI works, should they attract copyright. Further, once matter is understood as a thing or an object, it creates a dualist understanding of matter, where differences become essential in determining legal conceptualisation. Arguably, this creates a distraction from fully understanding AI, as questions of ‘ownership,’ ‘authorship,’ ‘human/non-human’ contributions take the forefront of many debates, without fully understanding AI. This paper predicates that an ontological theory such as new materialism should be adopted to better understand AI. This can be done by focusing on the materials and material effects of an AI work, so the processes involved in such a work become more visible. It should be noted that AI matter would include all matter including computer programs. This paper takes a forensic gaze into the legal materials that attempts to give legal meaning to computer matter. An examination of the case materials and the use of computer experts is conducted, providing an ontological understanding of computer matter, including AI, for copyright purposes.

Gaetano Dimita & Michaela MacDonald (Queen Mary University of London), We lost everything – Only video games can save us

We have put our faith into technology to deliver a better tomorrow. Yet, the recent scandals revealed the many vulnerabilities and exploits, threatening the very fabric of our society. Increasing privatisation of governmental functions and democracy has led to parallel yet disconnected modes of public and private governance. As governments are racing to save the remnants of democracy, or undermine it further, we are finding ourselves very much in the dark and at the mercy of private companies.

We advocate for virtual property, digital exhaustion, user rights and data ownership to safeguard our rights, freedoms, and agency. We need virtual property and digital exhaustion to free ourselves from platform dominance. We need user rights to counterbalance the expansion of scope and application of copyright, without a parallel evolution of limitations and exceptions, the ubiquitous presence of End User License Agreements, and the abuse of technological protection measures, all of which have disenfranchised the ordinary citizens. We need data ownership in order to return to being customers rather than the products being sold.

BIOS

Kwaku Adu-Amankwa is a PhD candidate based at the University of Strathclyde where he researches into the complex security relationships associated with intellectual property of additive manufacturing applications within digitally-enabled supply chains. Additionally, he has international industry experience from various roles held within oil & gas, design & manufacturing, and production engineering sectors across Ghana, USA, UK, China & Nigeria.

David Felipe Alvarez-Amézquita is an Assistant Professor and Researcher at the University of Tolima, Colombia. Leader of the Research Group on Private Law. Ph.D. in Laws / The University of Nottingham. Publications, research and teaching experience in the fields of Intellectual Property Law, History of Law, Law and Methods in various Colombian universities. Legal advisor on copyright for CERLALC-UNESCO and former Head of the Register Office at the National Copyright Directorate in Colombia. Law Degree and Magister Degree on History (Meritorious) at the National University of Colombia; LLM (Specialization) on Intellectual Property Law at the Externado University.

Lucio Casalini holds a Ph.D. in Private Law (cum laude) at Sapienza University, with a thesis on Commons and Intellectual Property Law. Master’s Degree in European Private Law (cum laude) at Sapienza University. Member of JODI and ESSENCE, two groups of research in the field of Digital Innovation and Finance Sustainability, established at Sapienza University. Author of many scientific articles. Honorary Research Fellow at University of Salento. Co-Director at Italian Scientific Journal specialized in Banking and Financial Law. Highly qualified Lawyer.

Daniela Caterino is Full Professor of Business Law and IP Law in the degree course of Law at Jonian Dept., University of Bari “Aldo Moro”. She  is also member of  Italian Banking and Financial Ombudsman and author of two books and several scientific papers on antitrust law, corporate governance, media law.   

Israel Cedillo Lazcano. Israel is a PhD candidate at the University of Edinburgh (Edinburgh Law School) and Full-Time Lecturer/Researcher in Banking and IP/IT Law at Universidad de las Américas Puebla (UDLAP) (Mexico). He obtained his Bachelor´s degree in International Law in 2009, a Master´s degree in Anthropologic Studies of Mexico in 2012 and a Master´s degree in Governance and Globalization in 2014, all of them from UDLAP). His academic research is focused on intellectual property, Artificial Intelligence, financial Innovation and the legal nature of money.

Jacopo Ciani is a post-doc researcher at the University of Turin and Adjunct Professor at ESCP Business School, Turin Campus (Italy). His research focuses on IP law in the digital environment. He is currently publishing his first monography on the Public Domain in the Information Society. 

Angela Daly, PhD is a socio-legal scholar of the regulation of new technologies based at the University of Strathclyde where she co-directs the Strathclyde Centre for Internet Law & Policy. She has been working on legal aspects of 3D printing for most of the last decade and is the author of Socio-Legal Aspects of the 3D Printing Revolution (Palgrave 2016).

Matthew David, PhD is Associate Professor in the Department of Sociology at Durham University. His research is primarily in the area of free-sharing online and its relationship to alternatives to business and alternative business models. His most recent book is David, M. (2021). [Forthcoming] Living in a Vacuum? Fabrications of regulation and deregulation in global network capitalism.

Giovanni De Gregorio is postdoctoral researcher working with the Programme in Comparative Media Law and Policy at the Centre for Socio-Legal Studies at the University of Oxford. Within the framework of the ERC ConflictNet project, his research focuses on content moderation and artificial intelligence; hate speech and disinformation in conflicts; digital policy in the global south. His research interest deals with digital constitutionalism, human rights, Internet law, privacy and data protection law.

Janice Denoncourt (BA McGill, LLB UWA, PhD Nottingham) is Associate Professor at Nottingham Law School, UK where her research involves multi-disciplinary issues related to intangibles, intellectual property, finance, corporate governance, data protection, entrepreneurship, and emerging technology law from a corporate governance perspective. Her monograph Intellectual Property, Finance and Corporate Governance (2018) addresses the lack of transparency and disclosure of intangibles and IPRs in corporate reports and the impact on shareholders and other stakeholders.  Before turning to academe, Jan was a corporate lawyer and Head of IP with Minter Ellison Perth, in-house counsel and alternate director for an ASX-listed company and Head of Legal Affairs for a large UK private company.  Janice founded and leads Nottingham Law School’s IP Research Group and is a Director of the IP Awareness Network, a London-based think tank.

Gaetano Dimita, PhD is a Senior Lecturer in International Intellectual Property Law, Centre for Commercial Law Studies, Queen Mary University of London. Gaetano is an Interactive Entertainment and Intellectual Property Law scholar and Video Games enthusiast. He is the editor-in-chief of the Interactive Entertainment Law Review, Edward Elgar, and the organiser of the ‘More Than Just a Game’ conference series.

Joshua A.T. Fairfield is William Donald Bain Family Professor of Law at Washington & Lee School of Law. He is an internationally recognized law and technology scholar, specializing in digital property, electronic contract, big data privacy, and virtual communities. He has written on the law and regulation of e-commerce and online contracts and on the application of standard economic models to virtual environments. Professor Fairfield’s current research focuses on big data privacy models and the next generation of legal applications for cryptocurrencies. His articles on protecting consumer interests in an age of mass-market consumer contracting regularly appear in top law and law-and-technology journals, and policy pieces on consumer protection and technology have appeared in the New York Times, Forbes, and the Financial Times, among other outlets.

Massimo Farina PhD – Adjunct Professor of “Computer Law” and “Digital Forensics” at DIEE – University of Cagliari. – Data Protecion Officer at  University of Cagliari. – Jean Monnet Key Staff of EU Data Governance, Cybersecurity and Digital Fundamental Rights (University of Milano) – Coordinator of ICT4Law&Forensics – Founder of DirICTo.

Tommaso Fia is a PhD Researcher at the Department of Law of the European University Institute (EUI) of Florence. His research interests encompass data governance, data ownership, and the commons. He is one of the coordinators of the Information Society Working Group of the EUI and collaborates on the EUI project ‘Unowned Personal Data, Interoperable EU Borders and Transitioning Rights’.

Erica Fraser B.Eng, LLB, LLM: Canadian Intellectual Property Office. Interest in implications of emerging technologies on intellectual property law.

Simon Geiregat is a Postdoc fellow of FWO (Research Foundation Flanders) at Ghent University Faculty of Law & Criminology (Belgium). Fellow (Stipendiat) at Max Planck Institute for Innovation and Competition. PhD in Law (Ghent 2019), thesis on copyright exhaustion and digital distribution. LL.M of IP & IT Law (KU Leuven 2017). ALAI and ELI member.

Sevra G. Guzel is a PhD Student at the Hertfordshire Law School. Her project investigates the online enforcement of copyright and the effects of this enforcement on fundamental rights. This project has been awarded with Hertfordshire Law School PhD Scholarship Award and been presented in various reputable international conferences. 

Vincenzo Iaia is Ph.D. candidate in Law and Business at Luiss University (Rome) and research fellow in Business and IP Law at the University of Bari. He is also Of Counsel at the law firm Akran Intellectual Property and junior researcher at BILL (Blockchain, artificial intelligence and digital transformation Law Lab).

Georgia Jenkins is a Herchel Smith Doctoral Researcher at Queen Mary University of London. She holds an LLM in IP Law with distinction from Queen Mary and an LLB from the University of Western Australia. Alongside her research, Georgia teaches copyright and trade mark law at Queen Mary.

Ignas Kalpokas (PhD, University of Nottingham) is Associate Professor at Vytautas Magnus University and LCC International University. His research encompasses the social impact of AI, media theory, political communication, and post-truth. He is the author of 3 books and nearly 30 journal articles and book chapters.

Julija Kalpokiene is a PhD candidate at the Faculty of Law of Vytautas Magnus University (Lithuania). Julija’s research interests concentrate on technological innovation, in particular artificial intelligence, and its regulation. She teaches a technology law course and has over 6 years’ experience of working in legal practice.

Alkistis Kympari holds a PhD in law from the University of Nantes, France (October 2020). Her PhD thesis entitled “Author’s rights and commercially unavailable works” examines solutions that facilitate uses of orphan and out-of-commerce works. Alkistis is member of the Institut de recherche de droit privé of the University of Nantes, and an attorney-at-law registered at the Bar of Athens and specializing in information law. Her research interests include digital rights management technologies and innovation policies in the digital age.

Christof Koolen is a Ph.D. researcher at the KU Leuven, Faculty of Law (Institute for Consumer, Competition & Market). He currently prepares a Ph.D. on the Internet of Things, focusing on the implications of smart devices on the internal market and, more specifically, on EU consumer law. Before starting with his doctorate, Christof interned at various law firms in Brussels and he provided legal advice on IP law issues to start-ups in the high-tech industry. He graduated with distinction from the KU Leuven (Belgium) in 2017 and obtained a Magister Juris degree from the University of Oxford (United Kingdom) in 2018.

Michaela MacDonald, PhD is Lecturer in Management, School of Electronic Engineering and Computer Science, Queen Mary University of London. Michaela teaches and researches on the impact of laws, norms and environmental constraints on users’ behaviour and interactions in Cyberspace. She is also the editor of the Interactive Entertainment Law Review, Edward Elgar, and a consultant at Moorcrofts LLP. 

João Carlos Magalhães is a senior researcher at the Alexander von Humboldt Institute for Internet and Society (Germany), and an incoming assistant professor in media and democracy at the University of Groningen (Netherlands). His research concerns the relationship between datafication, power and ethics, and has been published in journals such as International Journal of CommunicationSocial Media + Society and Journalism. João holds a PhD from the London School of Economics and Political Science.

Jacopo Menghini graduated in Law at the University of Turin in 2019. During the same year, he started a PhD at the University of Bologna (LAST-JD RIoE). Since his thesis, he has been interested in the topic of online parody and how it can affectthe digital environment.

Johan David Michels is a Researcher and PhD candidate at the Microsoft-funded Cloud Legal Project at Queen Mary University of London, and guest teacher at the London School of Economics. He focuses on questions of commercial law (such as contract, IP, and property) and IT, including cloud services, blockchain, and cybersecurity.

Lillà Montagnani, PhD is an associate professor of Commercial Law at Bocconi University, where she is also the Director of the LLM in European Business and Social Law (EBSL). She is currently a Visiting Professor at Peking University School of Transnational Law and a Transatlantic Technology Law Forum Fellow at Stanford Law School.

Aleksandra Nowicka is a First Class LLM and a Distinction GDL graduate, currently conducting her doctoral research at the University of Westminster, Westminster School of Law. She is a legal professional, and previously held a position of a judicial assistant in the civil appeal division.

Karen O’Neill is a doctoral researcher in copyright law at the University of Birmingham. Her research examines the ontological nature of computer programs and computer-generated works for copyright purposes. She is an experienced law lecturer, currently working at the Open University, teaching both contract and tort law.

Julio C. Padilla Herrera is Ph.D. candidate at the Los Andes University (School of Law, Bogotá) where he also obtained his Master’s degree in Law (LL.M). He is professor/researcher at Los Libertadores University. He is a recipient of the Colciencias Fellowship and his thesis is about the history of the link between property law and works of art in Colombia (1886-1982).

Alessia Palladino Phd “Humanities and Technologies”, Suor Orsola Benincasa University of Naples. – MSt Candidate European Public Law Organization (EPLO), Athens Greece. – Teaching Assistant “Computer Law” at DIEE, University of Cagliari.

Argyri Panezi is an Assistant Professor at IE Law School where she teaches contracts, copyright law, and principles of LegalTech. She is also a research fellow at the Digital Civil Society Lab at Stanford University, where she explores the notion of critical digital infrastructure with a particular focus on open source software. She holds a Ph.D. from the European University Institute and an LL.M. from Harvard Law School.  

João Pedro Quintais is Assistant Professor at the Institute for Information Law (IViR), University of Amsterdam. He is also a member of the European Copyright Society and managing editor of the Kluwer Copyright Blog.  His current research includes leading a Work Package on copyright content moderation and its impact on access to culture for the Horizon 2020 project ‘reCreating Europe’. NWO VENI-funded project on “Responsible Algorithms: How to Safeguard Freedom of Expression Online”.

Marco Ricolfi is Co-Founder and Co-Director of the Nexa Center for Internet & Society. He received his degree at the Faculty of Law of the University of Torino in 1974. Master of Laws, Yale Law School, 1975-1976. He is currently a professor of Intellectual Property Law at Turin Law School, where he also teaches Business Law. In 2007, Prof. Ricolfi was a member of the “High-Level Expert Group on the European Digital Libraries” created by the European Commission. He is currently a member of the European Copyright Society. Among his works The Internet of Things and the Ages of Antitrust, in Concorrenza e mercato 2017 (forthcoming); The new paradigm of creativity and innovation and its corollaries for the law of obligations, in (eds) Peter Drahos-Gustavo Ghidini and Hanss Ullrich, Kritika: Essays on Intellectual Property, Vol. I, Edward Elgar, Celtenham, 2015, 134-205 Trattato dei marchi. Diritto europeo e nazionale, Voll. 1 e 2, Giappichelli, Torino, 2015.

Anthony Rosborough is PhD Researcher at the European University Institute and Lecturer at the Schulich School of Law at Dalhousie University (Canada). His work focuses primarily on the ways in which intellectual property rights and technological design have been used to produce inequality by undermining individual agency and ownership.

Cristiana Sappa, PhD Italy and France, qualified attorney-at-law in France, is Associate Professor of Business Law at Iéseg school of Management in Lille and Paris, where she teaches IP law, contacts and business law to different target publics. Her research field is related to IP law, with a particular attention to copyright and cultural heritage institutes, open data, public sector information and trade secrets. She is co-coordinator of a working group on Knowledge commons and AI within the interdisciplinary research group on AI lead by the French CNRS. She extensively work on public sector information, in particular leading two European thematic Networks and one interdisciplinary research project on this topic. 

Burkhard Schafer is Professor of Computational Legal Theory at the University of Edinburgh. His main field of interest is the interaction between law, science and computer technology from doctrinal, comparative and legal-theoretical perspectives. He is currently Co-I of DeCaDe and Creative Informatics.

Giulia Schneider, PhD in International Law & Economics at Bocconi University (Milan), is Postdoctoral Researcher in Private Comparative Law at Lider Lab, Sant’Anna School of Advanced Studies (Pisa), with a particular focus on data protection and intellectual property law in digital markets. She is lawyer and lecturer at University of Pisa in Business & Commercial Law.

Amy Thomas is a Research and Teaching Associate and PhD candidate at the University of Glasgow, School of Law/CREATe centre. Amy is currently researching her PhD thesis (CMS/CREATe collaborative scholarship) which examines the conceptualisation of the user in end user licensing agreements. She also has interests in copyright-related aspects of video games and eSports, and acts as the managing editor for the Copyright Evidence Wiki, collating empirical research relating to copyright.

Sean Thomas, PhD is a Reader at the York Law School. His research centres on issues of ownership, with a special interest in the intersections between goods and other forms of property (land, IP). His recent work has focused on embedded software and smart technologies, as well as circular economy. 

Alina Trapova is an Assistant Professor in Law and Autonomous Systems at the University of Nottingham. She completed her law degrees in the UK (University of Sheffield and QMUL). She has worked at the IFPI, the EUIPO and in private practice, focusing on trade mark and copyright law. Prior to her move to Nottingham, she was a researcher at Bocconi University, where she coordinated the LLM in European Business and Social Law.

Jozefien Vanherpe is a PhD researcher at Consumer Competition Market (CCM) and an affiliated research fellow with the Centre for IT and IP law (CiTiP), both at KU Leuven (Belgium). Jozefien studied law at KU Leuven and holds an LLM degree from the University of Cambridge.

Ying Ye is a PhD candidate at Durham University. Ying is also a member and programme coordinator of the Centre for Chinese Law and Policy of the faculty of law. Ying’s research interest focuses on AI and intellectual property law, especially the copyright system. 


[1] [2018] EWCA Civ 518, appealed as UKSC 2018/0090, referred to the CJEU as Case C-410/19 The Software Incubator.

[2] Sean Thomas, ‘Sale of Goods and Intellectual Property: Problems with Ownership’ (2014) Intellectual Property Forum 25; Sean Thomas, ‘Goods with embedded software: obligations under Section 12 of the Sale of Goods Act 1979’ (2012) 26 International Review of Law, Computers & Technology 165.

[3] Sean Thomas, ‘Law, Smart Technology, and Circular Economy: All Watched Over By Machines of Loving Grace?’ (2018) 10 (2) Law, Innovation & Technology 230.

[4] PST Energy 7 Shipping LLC v O W Bunker Malta Limited [2016] UKSC 23.

[5] Louise Gullifer, ‘The Vanishing Scope of the Sale of Goods Act 1979 in the Twenty-first Century’ in Charles Mitchell and Stephen Watterson (eds), The World of Maritime and Commercial Law: Essays in Honour of Francis Rose (Hart Publishing 2020) 165.

[6] Aaron Perzanowski and Chris Jay Hoofnagle, ‘What we buy when we buy now?’ (2017) 165 University of Pennsylvania Law Review. See results of The MediaShop Study where the authors undertook the first-ever empirical study of consumers’ perceptions of the marketing language used by digital media retailors by creating a fictitious Internet retail site, surveying nearly 1300 online consumers.

Death Becomes HE: The Rise of Lecture Capture in the Panopt(ic)o(n)

Can universities force teachers to record their lectures? Are the declarations of academic freedom a mere performance? Are Zombie professors the future?

I’ve answered these and other questions yesterday when I presented ‘Death Becomes HE: The Rise of Lecture Capture in the Panopt(ic)o(n)’ at the coolest – and most feline – conference ever: Gikii 2021: The Light at the End of the Vaccination Syringe Edition (hosted virtually by Maanstricht University, 13-14 September 2021).

You can watch the video here and download the slides below:

This research is co-authored with Giulia Priora, Bernd Justin Jütte, and Léo Pascault and will appear with the more serious title of “Capturing the Uncapturable: The Relationship between Universities and Copyright through the Lens of the Audio-Visual Lecture Capture Policies” in the forthcoming book “Art and Literature in Copyright Law: Protecting the Rights of Creators and Managers of Artistic and Literary Works” edited for Edward Elgar by Enrico Bonadio and Cristiana Sappa.

In January 2021, Concordia University student and trans activist Aaron Ansuini was in shock when he found out that one of the professors who was teaching him had died in 2019. The lectures could be delivered by the zombie professor thanks to his lectures being recorded before his passing. The unsettling fact that the dead can replace living faculty members is the perfect metaphor for what is happening across higher education during the COVID-19 pandemic.

In light of this, the rules on lecture capture represent a meaningful entry point to investigate strengths and weaknesses of the universities’ approach to the creation and use of protected content online. Through this lens, it is possible to reflect on both the underlying exacerbated power imbalance between universities and teachers, and the diverging approach towards copyright law across the European higher education landscape.

First, the paper considers whether the selected universities had a specific policy on lecture capture, and, if so, what it covered, and where the default rule sat in the continuum between opt-out and opt-in.

Then, it investigates issues of ownership of the lecture recordings, including the incorporated slides and other materials as well as the performance rights. Despite stark differences between the selected countries’ approach, the clear trend towards an increased expectation that teachers have to record their lectures epitomises the digital dispossession that is inherent to the platformisation of education. Nonetheless, the oft-forgotten rights in performances can still play a role in pursuing a fairer balance between the competing interests at play.