Secret trigger words could make AI-powered robots dangerous
- AI-powered: 45 events in the last 90 days
- Previous: 3 days earlier · North Korea says it tested new AI-powered torpedo system
What happened
A team of computer scientists at Binghamton University has demonstrated that hidden trigger words inserted maliciously into LLMs could cause a robot to go off-script and perform dangerous actions—and they developed a defense.
Summary assembled by rule from the sources below