AI CHATBOTS CAN WRECK YOUR HEALTH

AI CHATBOTS often provide false health information — putting users' well-being at risk, a new study has found.

- 

Researchers at several institutions, including the University of South Australia, Flinders University, Harvard Medical School, University College London and the Warsaw University of Technology, found that chatbots developed by OpenAl, Google, Anthropic, Meta and X Corp can easily be programmed to deliver false medical and health information.

Using instructions available only to developers, the scientists programmed each Al system to produce incorrect responses to health queries. The answers were backed up with fabricated references from reputable sources to sound more credible.