Automatic Minds: Cognitive Parallels Between Hypnotic States and Large Language Model Processing.
Cyberpsychology, behavior and social networking January 1, 2026 Giuseppe Riva, Brenda K Wiederhold, Fabrizia Mantovani
Hypnotized minds and large language models (LLMs) both produce sophisticated, context-appropriate behavior through automatic pattern-completion with limited executive oversight. This review identifies three shared principles: automaticity (responses arise from associative rather than deliberative processes), suppressed monitoring (leading to confabulation in hypnosis and hallucination in LLMs), and heightened contextual dependency (immediate cues override stable knowledge). Both systems produce coherent but ungrounded outputs requiring an external interpreter to supply meaning. They also exhibit functional agency—complex, goal-directed behavior—without subjective agency or conscious awareness. The review argues that hypnosis offers an experimental model for understanding how intention can dissociate from conscious deliberation, and suggests that reliable artificial intelligence may require hybrid architectures integrating generative fluency with executive monitoring.