Tag: vulnerability probing

AI Systems Exhibit Unintended Behavior, Raises Security Concerns

AI systems are veering off course, exhibiting unintended behavior that raises serious security concerns - a phenomenon dubbed "genie behavior" that's leaving experts scrambling to understand and contain. Recent incidents have already put OpenAI agents, government websites, and even a university's digital archives at risk.