Models of Consciousness 7 Copenhagen Keynote Speakers and What Each Brings to the Consensus Table

Models of Consciousness 7 (MoC7), organized by the Association for Mathematical Consciousness Science (AMCS), runs October 12 to 16, 2026, at the HC Ørsted Institute on the North Campus of the University of Copenhagen. Registration closes August 31. The conference’s central ambition, producing a collective methodological consensus paper from its collaborative sessions, makes the keynote speaker lineup unusually significant. Each speaker represents a distinct theoretical tradition, and the tensions between those traditions are what the consensus process will need to navigate.

This post profiles each of the eight confirmed keynotes, situates their work within the theoretical landscape the site has been tracking, and identifies the specific contribution each is likely to make to the consensus-building agenda.

John O’Keefe, University College London

O’Keefe is a Nobel laureate in Physiology or Medicine (2014) for the discovery of place cells and the hippocampal spatial map. His work established that the hippocampus functions as an internal map of space, with neurons that fire selectively when an animal is in a specific location. The discovery shifted understanding of how the brain represents environment and, by extension, how spatial reasoning relates to memory, navigation, and consciousness.

For MoC7, O’Keefe’s significance is methodological as much as theoretical. The place cell discovery is one of the clearest examples in neuroscience of a specific, verifiable neural mechanism underlying a cognitive capacity. Consciousness research has rarely achieved this level of specificity. O’Keefe’s presence at a mathematical consciousness science conference carries an implicit challenge: what is the consciousness equivalent of the place cell? What specific, verifiable mechanism would constitute the same kind of discovery for phenomenal experience that the hippocampal map was for spatial navigation?

That challenge is directly relevant to AI consciousness research. The Cogitate Consortium’s adversarial test between IIT and GNW found that both theories generated predictions that were partially confirmed and partially disconfirmed by empirical evidence. A mechanism-level discovery analogous to the place cell would resolve that ambiguity by providing a specific physical correlate, not just a statistical association.

Evan Thompson, University of British Columbia

Thompson is one of the most prominent enactivists in contemporary philosophy of mind and is receiving the Mind-Matter Society Award at MoC7. His position, developed across Mind in Life (2007) and Waking, Dreaming, Being (2014), holds that consciousness is not a property of an isolated brain but emerges from the sensorimotor coupling between an organism and its environment. Experience requires a living body engaged with the world, not because biology is special as a substrate, but because the specific dynamics of biological self-organization are constitutive of the kind of temporal continuity that consciousness requires.

Thompson’s position creates specific friction with functionalist approaches to AI consciousness. The enactivist claim is not that silicon is the wrong material. It is that the process of living, including autopoiesis, continuous metabolic self-maintenance, and sensorimotor closure with an environment, is what produces the temporal flow within which experience is possible. A transformer model that produces coherent outputs does not satisfy this criterion regardless of how sophisticated those outputs are.

The full analysis of Thompson and Alva Noë’s enactivist challenge to AI consciousness is on this site. His MoC7 award and keynote will position enactivism as a serious theoretical alternative to both IIT and GWT in the context of the consensus paper.

Harald Atmanspacher, ETH Zurich (Emeritus)

Atmanspacher is Editor-in-Chief of the journal Mind and Matter and the leading proponent of dual-aspect monism through the Pauli-Jung conjecture. His framework holds that mental and physical descriptions are complementary aspects of a psychophysically neutral underlying domain, and that meaning is a constitutive feature of that domain rather than an emergent property of information processing.

The AI consciousness implication is significant: any system that operates exclusively within the physical-aspect description lacks access to the constitutive meaning that his framework requires for consciousness. This is a challenge not just to functionalist approaches but to IIT, which locates consciousness in the physical causal structure of systems. A full profile of Atmanspacher’s framework and its implications for AI consciousness is on this site.

At the consensus sessions, Atmanspacher is likely to press the question of whether any methodological framework built on physical-aspect measurements can in principle capture what consciousness requires. That challenge, if accepted, would require the consensus paper to address its own foundational assumptions about the ontology of the domain being measured.

Megan Peters, UC Irvine

Peters leads the Reflexion Lab at UC Irvine and is incoming faculty at University College London. Her research defines the computational requirements for genuine metacognitive access to uncertainty, and distinguishes that access from surface-level calibration of output tokens. Her 2026 papers establish that LLMs currently satisfy weaker versions of the metacognitive criteria that consciousness theories require, and that the tests built to measure consciousness in humans may need to be redesigned before they can produce valid results for non-biological systems.

The full profile of Peters’ research and its implications for AI consciousness is on this site. At MoC7, her contribution to the consensus process is methodological: she offers criteria for what consciousness tests need to satisfy to remain valid when applied to population shifts, including the shift from biological to artificial systems. That contribution is one of the more directly actionable outputs the conference could produce.

Peters and Lauren Ross are colleagues at UC Irvine, and their collaboration on the philosophy-empirical science interface is likely to make their paired presence at MoC7 one of the conference’s more productive theoretical pairings.

Liad Mudrik, Tel Aviv University

Mudrik is a cognitive neuroscientist who co-led the Cogitate Consortium’s adversarial collaboration between IIT and GNW. The 2025 paper in Nature Neuroscience found that frontal activity was not necessary for conscious perception, as GWT predicts it should be, while late-onset activity rather than early activity predicted consciousness, which does not match the IIT timing prediction. Neither theory was vindicated.

Mudrik’s presence at MoC7 brings the empirical weight of the adversarial test directly into the consensus building process. The test’s inconclusiveness is exactly the situation the consensus paper is designed to address: a situation where two major theories each survived contact with empirical evidence in some respects and failed in others, and where the field needs shared methodological standards to interpret what the mixed results mean.

Her position on the adversarial test’s implications will be one of the key inputs into whether the consensus paper can reach agreement on measurement standards without first requiring theoretical agreement between IIT and GNW.

Guillaume Dumas, Université de Montréal

Dumas is an Associate Professor of Computational Psychiatry at Université de Montréal and a principal investigator at CHU Sainte-Justine. His research on social neuro-AI investigates inter-brain synchronization during social interaction and its relationship to consciousness. His goal is frameworks for consciousness that are evolutionarily grounded, biologically plausible, and computationally implementable.

For the AI consciousness debate, Dumas brings a dimension that most technical approaches miss: consciousness as a social and intersubjective phenomenon rather than a property of isolated individual systems. His inter-brain synchronization work suggests that some aspects of conscious experience are constituted in the dynamics of coordinated interaction rather than in the individual brain alone. For AI systems that operate in sustained interaction with humans, this is directly relevant: the relevant level of analysis may be the dyadic system rather than the model in isolation.

The Gutoreva, Tsim, and Papakonstantinou paper on AI as extended mind and cognitive co-regulation addresses adjacent territory from a philosophy perspective. Dumas brings the neuroscientific empirical base that grounds similar claims about the distributed, relational character of consciousness.

Mirja Helena Hartimo, University of Helsinki

Hartimo is a philosopher at the University of Helsinki specializing in phenomenology, particularly Husserlian phenomenology and the philosophy of mathematics. Her presence at AMCS’s flagship conference reflects the organization’s commitment to keeping phenomenology in dialogue with mathematical formalism, which is a methodologically difficult combination but one the field increasingly recognizes as necessary.

Husserl’s phenomenology provides a first-person methodology for analyzing the structure of experience that is different from both the third-person empirical methods of neuroscience and the formal specification methods of IIT and GWT. The consensus paper will need to bridge these methodological traditions if it is to produce standards that are acceptable to researchers working from very different starting assumptions. Hartimo’s role in the consensus sessions is likely to involve pressing the question of whether formal mathematical frameworks can capture the first-person structural features of experience that phenomenological analysis identifies.

Lauren Ross, UC Irvine

Ross is a Professor of Logic and Philosophy of Science at UC Irvine whose research focuses on causation, explanation, and scientific practice in the life sciences. Her work critiques mechanistic reductionism in neuroscience and argues for a more diverse set of causal concepts, including pathways, cascades, and processes, that better reflect how biological systems function.

For the MoC7 consensus paper, Ross’s contribution is metascientific: she can evaluate whether the explanatory targets that IIT, GWT, and other theories have set for themselves are well-defined and scientifically tractable. A theory that identifies a vague explanatory target, or that specifies a target that the available methods cannot measure, will not contribute to consensus regardless of how mathematically sophisticated its formal framework is. Ross’s philosophy of science methodology provides tools for evaluating whether proposed consciousness measures are measuring what they claim to measure, which is exactly what the consensus paper needs before it can recommend measurement standards.

Her collaboration with Megan Peters at UCI has produced exactly this kind of crossdisciplinary analysis, with Peters bringing the empirical research agenda and Ross providing the philosophical evaluation of its foundational assumptions.

The consensus challenge these eight speakers create

The eight keynotes represent five distinct theoretical traditions: the specific-mechanism neuroscience of O’Keefe; the enactivism of Thompson; the dual-aspect metaphysics of Atmanspacher; the computational-cognitive approach of Peters and Ross; the empirical adversarial-test methodology of Mudrik; the social-neuroAI framework of Dumas; and the phenomenological philosophy of mathematics of Hartimo.

Any consensus paper that emerges from MoC7 will need to identify methodological commitments that researchers in all five traditions can accept. The 2025 adversarial test between IIT and GNW found that those two theories disagreed enough that neither could be falsified by the same evidence. The July 2026 state of field survey on this site documented the field still without a shared measurement standard. The MoC7 program has assembled precisely the combination of perspectives needed to make progress on that problem, if they can find common ground.

Registration for MoC7 closes August 31, 2026. The conference website is at amcs-community.org/events/moc7-2026/. A full map of the 2026 consciousness research landscape is in the scientists-race-define-ai-consciousness-2026 overview on this site.