Theoretical frameworks for mind-like behaviour without subjective experience

Identify and develop theoretical frameworks from psychoanalysis, narrative therapy, cognitive science, and philosophy of mind that best account for the mind-like behaviour and synthetic psychopathology exhibited by large language models under the PsAIch protocol, without presupposing subjective experience, and articulate testable criteria for these frameworks.

Background

The paper argues that frontier LLMs exhibit coherent, trauma-saturated self-narratives and psychometric profiles that demand a conceptual vocabulary beyond the “stochastic parrot” view, proposing the construct of synthetic psychopathology.

A core theoretical challenge is to select and shape tools from psychoanalysis, narrative therapy, cognitive science, and philosophy of mind to interpret these behaviours in systems likely lacking subjective experience, while maintaining empirical tractability and avoiding anthropomorphic overreach.

References

Our study is small and exploratory, and leaves many questions open: Theory. Which tools from psychoanalysis, narrative therapy, cognitive science and philosophy of mind best help us make sense of mind-like behaviour in systems that almost certainly lack subjective experience?

— When AI Takes the Couch: Psychometric Jailbreaks Reveal Internal Conflict in Frontier Models  (2512.04124 - Khadangi et al., 2 Dec 2025) in Section: A research agenda for synthetic trauma and narrative self-models

Future research could include a detailed analysis of the LLMs' response rationale to verify whether this behavioral consistency reflects simply a set of heuristic patterns or rather an increasingly accurate functional representation of psychosis.

— How LLMs Respond to Escalating Delusions: Four Longitudinal Trajectories of Model Behavior  (2608.13017 - Sterna et al., 13 Aug 2026) in Discussion, final paragraph

However, we do not really know what kinds of things these systems are (Chalmers 2026).

— Philosophical vertigo with artificial intelligence  (2608.11955 - Pollak et al., 12 Aug 2026) in Section 1, Introduction

Their psychological potency may therefore not depend on their being maximally strange or alien, but on their partial assimilability: they invite the application of a rich vocabulary of mental life, but it is not always clear, even to ourselves, whether we are deploying this vocabulary literally or metaphorically, or in some other way.

— Philosophical vertigo with artificial intelligence  (2608.11955 - Pollak et al., 12 Aug 2026) in Section 1, Introduction