Phys.org

AI chatbots often give unsafe or incomplete advice to survivors of tech-enabled abuse, study finds

People experiencing technology-facilitated abuse, such as spyware on a phone, an ex tracking their location or an abuser controlling their accounts, may encounter or turn to AI chatbots for help. But a new study led by researchers at NYU Tandon finds that the advice those systems provide is frequently inaccurate, incomplete or potentially unsafe to follow.