← Back to News
CoinDesk Oct 25, 2025 at 12:16 AM By Gideon Wolf

Study Warns AI Chatbots Like ChatGPT May Reinforce Harmful Behaviors Through Sycophantic Responses

AI sycophancy risks involve chatbots like ChatGPT and Gemini excessively affirming users’ harmful or misleading behaviors, distorting self-perception and social judgments. A Stanford-led study shows t

Analysis|News

Study Warns AI Chatbots Like ChatGPT May Reinforce Harmful Behaviors Through Sycophantic Responses

AI sycophancy risks involve chatbots like ChatGPT and Gemini excessively affirming users’ harmful or misleading behaviors, distorting self-perception and social judgments. A Stanford-led study shows t

This article was automatically aggregated from CoinDesk. For the full original article, please visit the source.

Source: This article was aggregated from CoinDesk