Cooperative AI: challenges and future directions
In today’s increasingly digitalized society, interactions between humans and artificial agents are becoming ever more frequent and complex. These interactions have the potential to significantly enhance human experiences, offering unprecedented opportunities for innovation and efficiency. However, they can also lead to frustration and pose severe threats to human dignity and autonomy. As Artificial Intelligence (AI) agents take on roles as advisors, delegates, and cooperative partners, we face a critical need to understand how these roles are accepted within our emerging hybrid societies.
The current state of knowledge regarding the rules that govern human-AI interactions, and the moral, ethical, and social implications of these interactions, is still limited. One of the central challenges is to determine the key ingredients that AI agents need to possess—or be perceived to possess—to be accepted as integral parts of our socio-ecological environment. How do different types of AI agents, whether humanoid robots or large language models (LLMs), influence the dynamics of human-AI interactions? To what extent should AI follow human rules, and are such rules adequate to sustain cooperation in these hybrid environments?
This panel brings together scholars from fields intersecting psychology, economics, philosophy and computer science to explore these questions from a multidisciplinary and critical perspective. We aim to discuss some of the latest research on these topics, fostering scientific debates that will help us better understand the cooperative potential of AI and the socio-psychological implications of its integration into human society. This panel is an invitation to explore the frontier of cooperative AI, its challenges, and its opportunities, ensuring that we harness the power of AI in ways that respect and enhance human values and societal well-being.
Invited Panelists:
Jean Francois Bonnefon, Institute for Advanced Studies in Toulouse
Ophelia Deroy, Ludwig Maximilian University Munich
Donatella Donati, Università dell’Aquila
Niccolò Pescetelli, London Interdisciplinary School
Bio and contributions:
Donatella Donati is an Assistant Professor of Logic and Philosophy of Science at the University of L’Aquila. Her research focuses on analytic metaphysics and the philosophy of artificial intelligence, with particular focus on the concepts of trust, trustworthiness, and moral agency. She is actively involved in multidisciplinary research teams dedicated to the ethical implications of AI.
Jean-François Bonnefon is a CNRS Research Director, with affiliations to the Toulouse School of Economics, where he chairs the Department of Social and Behavioral Sciences; the Toulouse School of Management; the Institute for Advanced Study in Toulouse of which he is the director; and the Artificial and Natural Intelligence Toulouse Institute where he is chair of moral Artificial Intelligence. A cognitive psychologist by training, he has held research positions across computer science, psychology, and economics departments — and his 100+ publications often feature a combination of these three domains. He is an expert on moral preferences and decisions, especially in the context of advanced Artificial Intelligence, and is widely known for his contributions to the ethics of autonomous driving.
Niccolo Pescetelli is a senior behavioural and data scientist. He has a DPhil in experimental psychology from the University of Oxford. He previously held positions at the MIT Media Lab and the Max Planck Institute. Niccolo’s work in collective intelligence investigates how groups of people and machines can be smarter together than working alone. Applications of his research inform how to design better collaboration platforms and governance tools. Niccolo is also the co-founder and Chief Scientist of PSi, a collective intelligence platform to host and analyse online conversations among hundreds of people.
Niccolo researches collective intelligence and group decision-making in humans and machines. His work focuses on understanding and designing collaborative systems blending human and machine intelligence. Niccolo is the founder and chief scientist of PSi, an online platform to host and analyse real-time online conversations at massive scales. His research has been published by top peer-reviewed journals and was featured on the international press (BBC, Business Insider, Forbes, El Pais).
Ophelia Deroy is a Professor of Philosophy and Neuroscience at Ludwig Maximilian University in Munich, where she leads the Munich Interactive Intelligence Initiative—a growing network of 40 core members and collaborators. She is also an Associate Research Director at the Institute of Philosophy in London.
Her research centres on a fundamental question: How do perception and experience come together from different sources—whether across senses, perspectives, or between humans and artificial systems? And more importantly, what happens when they don’t fully align? Her distinctive approach combines philosophical analysis with empirical research, moving from behavioral experiments to computational and NLP methods.
On the theoretical side, she challenges the idea that different sources of information ever completely integrate. Instead, she has developed new models to explain why interactions between sensory inputs, viewpoints, or even agents—whether human or AI—sometimes improve understanding and decision-making but, at other times, introduce ambiguity or bias.
On the applied side, she studies the mechanisms behind these interactions in perception, aesthetics, and decision-making, expanding into AI ethics and human-AI collaboration. Her work has contributed to major European and industry-led initiatives on AI, focusing on algorithmic decision-making, AI trust and uncertainty, and the non-anthropomorphic communication of AI systems. She served on expert board for the EU Parliament on Science and Technology, shaping discussions on the governance of AI, hybrid human-AI interactions, and the perceived politicisation of science and technological research.
Panel organizers
Eugenia Polizzi, Institute for Cognitive Sciences and Technologies, CNR
Costanza Alfieri, Department of Information Engineering, Computer Science and Mathematics, University of L’Aquila