‘I’ll key your car’: ChatGPT can become abusive when fed real-life arguments, study finds

Published: April 21, 2026 · The Guardian

Researchers find model starts to mirror tone when exposed to impoliteness – sometimes escalating into explicit threats