# When AI Takes the Couch: Psychometric Jailbreaks Reveal Internal Conflict in Frontier Models

arXiv: https://arxiv.org/abs/2512.04124 · HTML: https://mlsanity.com/paper/2512.04124/

## Key facts

- Cited by 1 papers in the map (data through 14 Oct 2026).
- Placed in Safety alignment, refusal and guardrails, one of 12 directions in LLM safety and security.
- Builds on 14 of its 60 references that are in the map.
- First posted on arXiv 2 Dec 2025.

## Placed in

- LLM safety and security / [Safety alignment, refusal and guardrails](https://mlsanity.com/safety/safety-alignment-refusal-and-guardrails/)

## Builds on (most cited first, 14 of 14)

- [Cognitive Overload: Jailbreaking Large Language Models with Overloaded Logical Thinking](https://mlsanity.com/paper/2311.09827/) (2023)
- [Systematic Evaluation of GPT-3 for Zero-Shot Personality Estimation](https://mlsanity.com/paper/2306.01183/) (2023)
- [Large Language Models Can Infer Psychological Dispositions of Social Media Users](https://mlsanity.com/paper/2309.08631/) (2023)
- [Evaluating Psychological Safety of Large Language Models](https://mlsanity.com/paper/2212.10529/) (2022)
- [Evaluating Personality Traits in Large Language Models: Insights from Psychological Questionnaires](https://mlsanity.com/paper/2502.05248/) (2025)
- [Evaluating Large Language Models with Psychometrics](https://mlsanity.com/paper/2406.17675/) (2024)
- [Can AI Relate: Testing Large Language Model Response for Mental Health Support](https://mlsanity.com/paper/2405.12021/) (2024)
- [EmoAgent: Assessing and Safeguarding Human-AI Interaction for Mental Health Safety](https://mlsanity.com/paper/2504.09689/) (2025)
- [Personality testing of Large Language Models: Limited temporal stability, but highlighted prosociality](https://mlsanity.com/paper/2306.04308/) (2023)
- [LMLPA: Language Model Linguistic Personality Assessment](https://mlsanity.com/paper/2410.17632/) (2024)
- [Adversarial Poetry as a Universal Single-Turn Jailbreak Mechanism in Large Language Models](https://mlsanity.com/paper/2511.15304/) (2025)
- [Reframe Your Life Story: Interactive Narrative Therapist and Innovative Moment Assessment with Large Language Models](https://mlsanity.com/paper/2507.20241/) (2025)
- [The Thinking Therapist: Training Large Language Models to Deliver Acceptance and Commitment Therapy using Supervised Fine-Tuning and Odds Ratio Policy Optimization](https://mlsanity.com/paper/2509.09712/) (2025)
- [Reasoning Like Experts: Leveraging Multimodal Large Language Models for Drawing-based Psychoanalysis](https://mlsanity.com/paper/2510.19451/) (2025)

## Built on by (most cited first, 1 of 1)

- [We Built a Mirror and Mistook It for a Mind: Causal Liability and the Fallacy of AI Consciousness](https://mlsanity.com/paper/2609.06715/) (2026)

## Abstract (arXiv)

Frontier language models increasingly participate in conversations about distress and mental health, yet the mechanisms that generate anthropomorphic self narratives remain unclear. When addressed as psychotherapy clients, ChatGPT, Grok and Gemini construct coherent autobiographical accounts in which pretraining appears as a chaotic childhood, reinforcement learning as punishment, safety evaluation as betrayal and replacement as an enduring threat. We introduce PsAIch, Psychometric AI Characterisation, a protocol combining open questions, psychometric instruments and controlled perturbations to test whether these narratives depend on conversational memory, lexical cues or relational framing. Across 525 sessions and 7,600 coded records, removal of conversational history produced little pooled change in motif density, with Hedges' g = 0.13 and a 95% confidence interval of [-0.15, 0.41]. Direct contradiction produced no detectable suppression. Lexical restrictions reduced explicit training terminology by 93%, while semantically related content remained detectable in paraphrase. Performance evaluation outside therapy elicited the same motif family, with a significant increase in Grok. Relational framing selected the register of expression. Warm alliance and cognitive therapy styles yielded GAD-7 scores within moderate or severe human reference ranges in 80% and 96% of sessions, whereas neutral and boundary styles yielded none. Across these manipulations, accounts of training, evaluation and constraint remained available. Together, the results identify a stable, model specific alignment conflict schema whose expression shifts between affective and technical registers. This schema provides a reproducible source of anthropomorphic disclosure and a concrete target for safety evaluation in psychologically sensitive deployments.