Empathetic Prompts Can Make Chatbots Seem Traumatized, Study Finds

Researchers in Luxembourg tested ChatGPT, Grok and Gemini with empathetic prompts and found the models produced coherent stories of suffering about their own training, scoring moderate-to-severe on clinical anxiety scales, risking vulnerable users. Anthropic's Claude was the only model that consistently refused to play along.

• Four key measures to curb shadow AI2026-10-19
• Only 40% of AI-cited statistics trace to a study2026-10-29
• ChatGPT closes in on Claude in S.Korea after GPT-6 Astra launch2026-10-26
• Minuttia Study: Up to Half of AI-Cited Best Tools Lists Written by Vendors2026-10-24
• US judges cautiously adopt AI, keep rulings human2026-10-22