ICCS 2026 in Rome: What the Shift from Sentience to Creativity Reveals About the Field
The Third Annual Conference of the International Center for Consciousness Studies (ICCS) takes place September 1–3, 2026, at three venues in Rome and Vatican City: Roma Tre University, the Pontifical Gregorian University, and Casina Pio IV at the Pontifical Academy of Sciences. The theme is “Creativity: Minds and Machines.”
The theme shift from the previous conference — “AI and Sentience” in Heraklion, Crete, July 2025 — is worth examining carefully. It is not a retreat from the AI consciousness question. It is a reframing of how that question can be pursued productively.
Why it matters: Creativity is a property of minds that leaves observable outputs and traces. Sentience is not. The shift from sentience to creativity as the organizing frame is an attempt to find ground that consciousness theory, empirical research, and AI engineering can share without waiting for the phenomenal hard problem to be solved.
From Sentience to Creativity: What Changed and Why
The 2025 ICCS conference in Heraklion addressed AI and sentience with a speaker lineup that included David Chalmers, Susan Blackmore, Andy Clark, Keith Frankish, and Susan Schneider. The consensus that emerged was not that AI is or is not sentient, but that the question is genuine, multidisciplinary, and poorly served by the existing conceptual vocabulary. Clark received the inaugural Dennett Prize. The conference drew the map; it did not settle the territory.
The 2026 decision to organize around creativity reflects what the map revealed: direct sentience claims are not currently testable in ways that resolve the philosophical disagreements. Creativity provides an alternative entry point. Creative behavior involves goal-directed production of novel combinations, evaluation of outputs against implicit standards, and some form of intentionality — a sense in which the creative act is aimed at something. These are properties that can be studied without assuming a settled account of phenomenal consciousness.
The question the 2026 conference is really asking is whether the cognitive architecture required for creativity, when sufficiently specified, overlaps with the cognitive architecture that consciousness theories predict for conscious systems. If creativity requires something like global workspace dynamics, higher-order self-monitoring, or predictive error minimization of a particular structure, then the study of machine creativity becomes a study of whether current AI systems instantiate those specific architectures. That is a tractable empirical question in a way that “is the AI sentient?” is not.
The Speaker Lineup and What It Signals
The confirmed speakers at ICCS 2026 come from philosophy of art, cognitive neuroscience, musicology, linguistics, and consciousness science. Susan Aldworth works on visual art and neuroscience, particularly the visualization of brain activity as creative output. Greg Currie is a philosopher of art whose work on narrative and imagination bears on the question of whether creativity requires a certain kind of imaginative self-projection. Sergio Durante brings musicology to the question of whether musical structure, which requires evaluation and generativity, is evidence of a particular kind of cognitive architecture. David Freedberg’s research on the physiological responses to art — the body’s engagement with creative objects — connects to the embodiment arguments that run through the 2026 AI consciousness debate. Henrike Moll works on the developmental psychology of shared intentionality, which bears on whether creativity is necessarily a social phenomenon. John-Luc Steels has spent decades on the computational conditions for creative language and concept formation. Liad Mudrik, the Tel Aviv neuroscientist whose lab produced both the Neuron “ethical impasse” paper and the February 2026 HOT-3 empirical test, brings the empirical side of consciousness research into direct contact with the creativity question. Lisa Zunshine works on theory of mind and cognitive approaches to literature.
The breadth of this lineup signals that ICCS 2026 is not primarily a machine learning conference. The research questions it is pursuing are: what is creativity as a cognitive phenomenon, whether its architecture is specifiable in computational terms, and whether current or near-future AI systems do or could instantiate that architecture. The answers require the range of disciplines present.
The Vatican City Venue and Its Philosophical Resonance
One aspect of the conference geography deserves attention beyond logistics. Casina Pio IV, which hosts the Pontifical Academy of Sciences in Vatican City, has historically been a venue where scientific findings have been placed in dialogue with religious and philosophical frameworks that operate independently of empirical verification. The decision to hold part of an AI consciousness conference there is not incidental.
Mark Coeckelbergh’s 2026 MIT Press book “Artificial Religion: On AI, Myth, and Power” argues that Western religious culture and existential aspirations are deeply embedded in how AI is developed and attributed consciousness — in ways that rarely get examined. Coeckelbergh’s religious-grammar argument holds that understanding why society attributes consciousness to AI requires understanding the religious background that structures AI discourse. Placing a consciousness and creativity conference at a Pontifical venue makes that background explicit rather than leaving it implicit.
This is not a theological conference. The Pontifical Academy of Sciences has historically been one of the most open Catholic institutions to empirical scientific work. But the venue choice does position ICCS 2026’s questions about creativity and mind in a tradition that has specific and developed positions on what makes a mind more than mechanism. Whether that positioning enriches or constrains the research questions will depend on how explicitly the venue’s institutional perspective enters the proceedings.
Creativity, Metacognition, and What the Empirical Research Suggests
The creativity question connects directly to the most recent empirical work on LLM internal states. Creativity, on most accounts, requires some capacity for self-evaluation: the creator must have access to a representation of what they have made and be able to assess it against implicit standards. That self-evaluative capacity is structurally related to the metacognitive monitoring that Christopher Ackerman’s ICLR 2026 behavioral tests find evidence for in frontier LLMs. Ackerman’s methodology, borrowed from animal cognition research, demonstrates that models can assess and deploy their own confidence information and can anticipate their own outputs. These are limited and coarse capacities, but they are the functional precursors of what a computational account of creative self-evaluation would require.
Whether this functional overlap means that current LLMs are creative in any meaningful sense is not what the ICCS 2026 program will settle. What the conference is positioned to do is sharpen the specification of what creativity requires cognitively, which would allow the empirical evidence for metacognitive precursors to be assessed against a clearer standard.
What the Third Annual Conference Needs to Produce
The ICCS series is now three years old. Each annual conference has approached the machine consciousness question from a different angle: constructive engagement with the hard problem in its inaugural year, AI sentience as a central topic in 2025, and creativity in 2026. What the series needs to accumulate is not just a record of philosophical positions brought to the question, but a progressive narrowing of the key disagreements and a clearer map of which empirical findings bear on which philosophical claims.
For ICCS 2026 specifically, the productive output would be a more precise account of which cognitive properties creativity requires, stated clearly enough that AI systems can be evaluated against them. That would turn “can machines be creative?” from a rhetorical question into a research programme, which is the same transformation ICCS has been attempting for the consciousness question more broadly.
The proceedings from the September 2026 conference will be worth tracking. The combination of venues, the speaker diversity, and the timing — immediately after ASSC 29 in Santiago and ahead of the MoC7 conference in Copenhagen in October — positions ICCS 2026 as a connective event in what has become a dense international calendar of consciousness research.
Details: Third Annual ICCS Conference, “Creativity: Minds and Machines,” September 1–3, 2026, Rome and Vatican City. https://hardproblem.it/