Hampshire College, Amherst, MA, USA
What Is It Like to Be a Bot? (AI and the Crack in the Mirror of Consciousness) — The relationship between consciousness and its material substrate has been theorized primarily as a question of instantiation: which physical processes are sufficient to generate subjective experience, and whether those processes can be reproduced in non-biological systems. The coupling between the signals by which interiority is recognized and the processes by which interiority is generated has been treated as a logical necessity rather than a historically contingent feature of particular substrate conditions. The LLM transformer invalidates this assumption through a specific mechanism formalized here as a semiotic phase transition: the reproduction of the signals of interiority below the threshold at which their generative processes become symbolically decomposable or introspectively verifiable. This transition exhibits a structural parallel to quantum accounts of state resolution in which continuous possibility resolves into discrete symbolic commitment only through an observer-dependent stabilization event. Substrate independence is not thereby refuted and the question is no longer understood to be whether an artificial system can instantiate a mind, but what occurs when the external signature by which a mind becomes knowable can be generated without access to the substrate from which that signature historically emerged.
Forest Mars is an independent researcher and writer working at the intersection where AI systems research meets philosophy of mind and consciousness studies. After his undergraduate work in Cybernetics he has spent his career building advanced computational systems, including work on frontier AI systems. His current research examines how changes in computational substrates alter the relationship between cognition, representation, and self-modeling, with particular focus on the implications of large language models for theories of consciousness and intelligence.
His recent work develops theoretical frameworks for understanding AI systems as a new threshold in the history of cognitive representation, exploring the relationship between symbolic emergence, machine intelligence, and the evidential basis by which cognition is recognized. He has presented his work at venues including the City University of New York, IBM, and the United Nations, and writes on artificial intelligence, cybernetics, and the future of cognition.