CogSci 2026 in Rio de Janeiro: What the Cognitive Science Society Brings to the AI Consciousness Debate
The 48th Annual Meeting of the Cognitive Science Society runs July 22 through 25, 2026, in Rio de Janeiro, Brazil, in hybrid format. Information on the program, keynotes, and registration is available at https://cognitivesciencesociety.org/cogsci-2026. The conference has added a dedicated workshop on the Cognitive Science of AI Alignment — new for 2026 — that treats the alignment problem as a cognitive science question rather than a technical optimization challenge.
CogSci is not a consciousness conference. It is not an AI safety conference. It is the primary annual gathering of researchers who study the mechanisms of mind across computational, behavioral, neural, and developmental perspectives, in humans, animals, and increasingly in artificial systems. That institutional position gives the conference a distinct angle on AI consciousness that none of the summer’s other major meetings provide.
What CogSci Is and What It Contributes
The Cognitive Science Society was founded in 1979 at a moment when cognitive psychology, AI, linguistics, philosophy of mind, and neuroscience were converging on a shared set of problems: how does the mind represent information, process it, and use it to produce behavior? The society is the institutional home of that convergence.
This means CogSci approaches AI consciousness from a different direction than the conferences that make AI consciousness their explicit subject. The Association for the Scientific Study of Consciousness (ASSC 29 meets in Santiago, June 30 through July 3) approaches AI consciousness from neuroscience and phenomenology, treating consciousness as a property of biological systems that may or may not extend to artificial ones. The AISB 2026 symposium at Sussex (July 1-2) approaches it through philosophy of mind and policy analysis. ASSC 29 produces experimental findings on neural correlates of consciousness; AISB produces philosophical framework and governance recommendations.
CogSci adds a third approach: how do cognitive mechanisms, the specific computational and representational processes that produce intelligent behavior, relate to the properties that consciousness researchers are trying to identify? This question has a long history within cognitive science that predates the current AI consciousness debate.
The computational theory of mind, which drove much of cognitive science’s early development, held that mental processes are computational in nature and that understanding the computational structure of cognition is the primary path to understanding the mind. That framework is in tension with the biological naturalism position (associated with John Searle and, in current AI consciousness debate, with Anil Seth’s controlled hallucination framework) that insists on substrate specificity. CogSci is the venue where this tension has been worked out over decades, and that historical depth is part of what it contributes to the 2026 AI consciousness discussion.
The Cognitive Science of AI Alignment Workshop
The new workshop for 2026 is analytically significant beyond its specific content. Its framing — treating alignment as a cognitive science problem — implies that the failure modes of AI alignment are failures of representation, reasoning, or generalization rather than failures of optimization or specification. That is a substantive theoretical commitment, and it has direct implications for AI consciousness research.
The standard technical framing of alignment asks: how do we specify objective functions that represent human values, and how do we optimize AI systems to pursue those functions reliably? The cognitive science framing asks a different question: what kinds of representations, reasoning processes, and generalization capacities does an aligned AI system need, and how do those differ from what misaligned systems have?
This reframing matters for consciousness research because the cognitive science questions about alignment are close to the questions that consciousness researchers are already asking about AI systems. Both ask about how AI systems represent the world, represent their own states, reason about goals and values, and generalize from training to novel contexts. The 2026 field survey of AI consciousness research documents a convergence between the mechanistic interpretability program (which asks what internal representations AI systems have) and the indicators framework (which asks which functional properties a conscious AI would exhibit). The cognitive science of alignment workshop sits at the same intersection.
Specifically: if an AI system has the kind of robust world-model, self-model, and value-representation that cognitive scientists think alignment requires, does that architectural profile overlap with the architectural profiles that consciousness theorists think consciousness requires? The workshop does not answer this directly, but it brings researchers from alignment and cognitive science into a common program that the consciousness-specific conferences are not structured to do.
Rio de Janeiro and the Global Distribution of Consciousness Research
The choice of Rio de Janeiro for CogSci 2026 follows ASSC 29 in Santiago. Two of the three major summer consciousness-adjacent conferences in 2026 are in South America, and both have hybrid formats that expand participation beyond the traditional North American and Western European academic centers.
This geographic distribution reflects a real change in the research community. Cognitive science and AI consciousness research are now conducted across institutions in South America, East Asia, and Africa that were largely absent from the field’s founding generation. The hybrid format is a structural acknowledgment that the questions CogSci addresses are not disciplinarily or geographically contained.
For AI consciousness research specifically, the Rio conference arrives in the same month as the AISB symposium at Sussex (July 1-2) and one week after ASSC 29 closes (July 3). The July 2026 conference cluster, ASSC 29, AISB, and CogSci within a four-week window, means the field is producing a dense concentration of presentations, critiques, and synthesis in a short period. Researchers engaging with the Butlin et al. indicators framework, the interpretability-based consciousness evidence, and the methodological critiques of both will have material from all three conferences to work with by mid-August.
The Persistent Question CogSci Brings Back
One tension that CogSci consistently surfaces is whether cognitive science’s standard approach, studying the mechanisms of cognition in terms of information processing, representation, and computation, is adequate for addressing consciousness at all.
The hard problem of consciousness, in David Chalmers’ formulation, is precisely the problem that survives the complete mechanistic description of cognitive processing. Explaining how a system processes information, represents states, and generates behavior is the “easy” problem (easy in that there is a principled methodology for addressing it, even if the details are hard). Explaining why that processing is accompanied by phenomenal experience is the hard problem, and it is not addressed by cognitive science’s standard methodology.
The AISB 2026 symposium at Sussex approaches this gap through philosophical argument and policy analysis. CogSci approaches it through the expansion of cognitive science’s methodology to include phenomenological and first-person methods — influenced by Varela’s neurophenomenology program, which is receiving dedicated attention at the July 4-5 satellite workshop at ASSC 29 in Santiago. The methodological question is the same across both venues: whether cognitive science can incorporate first-person evidence without losing scientific rigor.
Whether the 48th CogSci meeting advances this question, or whether the AI Alignment workshop produces insights that feed back into the consciousness debate, will depend on what is presented in Rio in July. The conference is worth tracking by researchers who want the cognitive science perspective on questions that are more typically addressed from philosophy, neuroscience, or AI research.