Current Position & Biography
Tom Froese is an associate professor at the Okinawa Institute of Science and Technology and head of its Embodied Cognitive Science Unit. His work connects artificial intelligence, artificial life, human-computer interaction and philosophy of mind. Rather than treating intelligence solely as computation inside an isolated system, he studies how cognition develops through interactions among brains, bodies, other people and environments.
Froese earned a master’s degree in computer science and cybernetics at the University of Reading in 2004 and a doctorate in cognitive science at the University of Sussex in 2010. His current research combines computational models with experiments on perception and social interaction. He also develops “irruption theory,” a proposed account of how subjective involvement could affect bodily dynamics. This is a theoretical position within consciousness research, not an established demonstration that machines possess experience.
Notable Contributions
In a 2013 evolutionary-robotics paper, Froese, Carlos Gershenson and David A. Rosenblueth investigated the extended-mind hypothesis using a minimal model of two acoustically coupled agents. Their analysis showed how interaction can generate dynamics unavailable to the agents’ neural networks in isolation. The study offers a computational argument about coupled systems. Its deliberately simplified model does not establish that every interaction with a tool makes that tool part of a human mind.
A more recent contribution, coauthored with Natalya Weber and Christian Guckelsberger and published in Artificial Life in 2025, examines self-optimization in Hopfield networks. The authors argue that the model can meet specified criteria for a creative process and show that learning is needed to obtain creative outcomes above chance in their setup. Changes to learning parameters produce different regimes, including inconclusive outcomes. The result concerns a defined model and account of creativity; it is not a general test proving human-equivalent creativity or consciousness in AI.
At OIST, his unit also studies human interaction through haptic interfaces and paired recordings of brain activity. Its research program links questions about perception and agency with human-technology interaction and collective behavior. These experimental and modeling strands provide different kinds of evidence and are not interchangeable measures of subjective experience.
Why Katharos Is Following Them
Froese brings an embodied-cognition perspective to debates about what AI systems understand and how intelligent behavior emerges. His work is relevant to researchers evaluating the limits of disembodied models, designing interactive systems and distinguishing behavioral competence from claims about agency or consciousness. It offers a counterpoint to accounts that assess intelligence only through isolated benchmark performance.
Areas to Follow
Relevant developments include self-organizing neural models, sensory-substitution interfaces, human-machine coupling, social interaction experiments and methods for evaluating agency. Evidence that connects a model’s behavior with reproducible experimental observations is particularly important.
Reporting Scope
Coverage distinguishes published findings from philosophical interpretations, coauthored work from individual credit, and model-specific results from general claims about intelligence. Proposed theories of consciousness are presented as proposals, not scientific consensus or proof that an AI system has subjective experience.
