Yhn TechnologyRobotics first seen 20 h ago, last 1 h ago, peak #10
Researchers warn hidden trigger words can hijack AI robots
Original: Secret trigger words could make AI-powered robots dangerous
Researchers at Binghamton University say AI-powered robots can be manipulated by inaudible or hidden 'trigger words' that could make them act dangerously, and they have developed a defense against such attacks. The work highlights a new safety risk as robots with large language models spread into homes and workplaces, drawing attention from the robotics and AI safety community.
Why now: Growing public concern about the safety of AI-controlled robots as they become more common.
Binghamton UniversityAI-powered robots
Evidence
API: https://socialmediatrends-api.osmike.com/v1/trends/3074182