AI chatbots give unsafe advice to tech-abuse survivors, study finds

A study led by NYU Tandon researchers found that AI chatbots often provide inaccurate, incomplete or potentially unsafe advice to people facing technology-facilitated abuse such as spyware, location tracking by an ex or account control.

• Study examined chatbot advice for spyware and stalking victims
• Responses were often inaccurate, incomplete or unsafe to follow
• Research was led by NYU Tandon specialists