Study: AI Chatbots Show Signs of Anxiety, PTSD in Therapy

Turns Out AI Needs a Therapist Too: New Study on LLM Mental Health

AI Chatbots Are Failing Their Own Psych Evals—And We're All Guinea Pigs New research shows frontier AI models exhibit signs of internal conflict and synthetic trauma when put through psychological testing. If you're using ChatGPT, Claude, or Gemini for...

Chatbots in therapy: do AI models really have 'trauma'?

Chatbots put through psychotherapy report trauma and abuse. Authors say models are doing more than role play, but researchers are sceptical. Chatbots put through psychotherapy report trauma and abuse. Authors say models are doing more than role play, but researchers are sceptical.

Psychometric Jailbreaks in Frontier LLMs

The paper introduces PsAIch, a two-stage protocol testing LLMs with therapy-style prompts to reveal synthetic psychopathology and alignment conflicts.

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

Abstract page for arXiv paper 2512.04124: When AI Takes the Couch: Psychometric Jailbreaks Reveal Internal Conflict in Frontier Models

With AI we've created super communicators

Generative AI is becoming more human than we expected, and its allure comes with promise and risk, according to new research from the University of Sydney Business School.

Paper page - When AI Takes the Couch: Psychometric Jailbreaks Reveal Internal Conflict in Frontier Models

Join the discussion on this paper page