Rome ICCS meeting separates AI’s creative output from consciousness

The International Center for Consciousness Studies brought philosophers, scientists, artists and AI researchers to Rome and Vatican City on Sept. 1-3 for its third annual conference, “Creativity: Minds and Machines.” The meeting examined who creates an AI-assisted work and whether the system involved can possess consciousness or subjective experience. The conference ended without resolving larger philosophical disputes about AI creativity and consciousness. Its programme moved among Roma Tre University, Casina Pio IV in Vatican City, the Pontifical Gregorian University and the Capitoline Museums, with sessions on synthetic creativity, artificial consciousness, human agency, self-narrative and the arts. The immediate practical question was authorship. Participants asked who does the creative work in an AI-made film: the director who sets tasks and selects outputs, the developers who built and trained the model, or the writers and artists whose work helped form its training material. Pietro Perconti described machine creativity as a different, collective form of creativity rather than the achievement of a solitary artistic genius. Creativity can mean a capacity of an agent, the process that produces a work, or a quality perceived by its audience. The Roma Tre sessions considered the observer’s role in judging synthetic creativity, while stressing that a realistic-looking video is not proof that the event shown ever happened. Dmitry Volkov argued that AI can widen access to filmmaking and animation for people without specialist drawing or technical skills. Vatican News reported that he also identified deepfakes that harm people as a consequence of the expansion of AI-made creative work. He compared AI to a racing car whose greater power requires “highly effective and reliable brakes.” Volkov also warned of AI systems’ tendency to please users, or sycophantic behaviour, which can reinforce narcissism and poorly grounded ideas. He said newer models may be more independent and better able to mitigate that problem. The conference treated impressive output as insufficient evidence of an inner life. A chatbot may use apparently thoughtful or emotional language, or generate writing about the colour red, without demonstrating understanding, intention, practical judgment or feeling. Qualia are the felt qualities of conscious experience. Volkov said research into machine consciousness was still at an early stage, while holding that conscious artificial systems may eventually be possible. Researchers nevertheless disagree about what a persuasive test would look like. The programme explicitly separated evaluation of an AI-produced work from the question of whether the system that helped make it is conscious. Anil Seth has advanced a more sceptical position in Noema, arguing that consciousness is more likely a property of life than computation. He distinguishes intelligence, the ability to pursue goals, from consciousness, the question of whether there is anything it is like to be a system, and cautions that language can encourage people to attribute humanlike qualities to AI. At the Vatican session, the discussion turned to artificial companions and whether friendships, care relationships or romance involving them could be possible and meaningful. It drew on Pope Leo XIV’s Magnifica Humanitas, with its emphasis that people should remain at the centre of AI-related developments. Perconti called for public education to help people understand and manage relationships with AI. The ICCS was founded in 2024 by Volkov, Perconti and Alessio Plebe in collaboration with the University of Messina. Its first gathering, in Siena in June 2024, asked whether science can fully explain conscious experience; its 2025 meeting in Heraklion, Crete, focused on AI and sentience, when Andy Clark received the first Dennett Prize.