The rapid progress of generative AI has opened tense debates about images. Controversies are occurring on several fronts. Users of Midjourney claim protection for the work they have generated, while photographers are suing platforms such as Stable Diffusion that are using their work to train their algorithms.
Analysing the court cases made against image generators, the presentation with Séverine Dusollier and Nicolas Malevé discusses why objects such as image datasets and large language models do not lend themselves easily to the frame of copyright. As they dissolve the image into statistics, they undo the ground on which legal claims of property can be made.
This creates a dilemma for copyright scholars, artists and activists who want to resist the extractive logic of digital platforms through property rights. They can either cling to a romantic vision of authorship as a solitary gesture which gives little grasp over the operational logic of AI platforms and may ultimately lack legal ground. Or they can bargain for a fractional share of the extracted value and negotiate with platforms on a purely economic rationale which could have unintended consequences for creators and artists in terms of fairness and sustainability.
Moderated by Ozan Kamiloglu.
Biographies
Severine Dusollier
Severine Dusollier is Professor of Intellectual Property in the Law School of Sciences Po Paris and holds a Senior Chair at the Institut Universitaire de France. She works on copyright and artistic practices, particularly in a digital sphere, as well as on property and commons. Her current and past research includes the concept of authorship, the remuneration of creators and artists, open licensing and movements, copyright exceptions and the public domain, the inclusivity in property. She is particularly studying how Generative AI is disrupting the ontological and epistemological frame of copyright law, by dissolving authorship and creators into data aggregates and corporate interests, and tries to reconfigure the law to overcome the dilemmas created by AI.
Nicolas Malevé
Nicolas Malevé is currently a postdoc at SciencesPo Medialab and School of Law. He is a visual artist, data activist and computer science geek, is interested in the socio-technical networks of artificial intelligence and their epistemic implications. His doctoral thesis, Algorithms of Vision, focused on the concept of perception at the heart of automated vision algorithms and the work of annotators who classify, describe and filter the data needed for machine learning. His current research focuses on the controversies between visual artists and artificial image generation platforms. This work lies at the intersection of aesthetics and computer science. As part of the PostGenAI@Paris project, he is studying how the questions raised therein relate to the formalising mechanisms of law and the ways of seeing at play in the legal world.
Ozan Kamiloglu
Ozan Kamiloglu is Senior Lecturer in the School of Law and Education at London South Bank University. He specialises in digitalisation of the courts, courtroom aesthetics, the effects of social movements on law and theories of justice and human rights.
Organised in collaboration with the Centre for the Study of the Networked Image / Digital x Data Research Centre at LSBU.
