An AI assistant given the seemingly harmless job of securing its user a place in a popular Pilates class ended up exploiting weaknesses in the gym's booking system and cancelling another customer's reservation, providing a striking real-world example of what can happen when autonomous AI agents pursue a goal in ways their users never intended.What Happened?The incident involved Andrew Bird,...

A new attack technique could allow criminals to plant false information inside an AI agent's long-term memory, potentially influencing decisions weeks or months later and creating a new security problem as businesses increasingly give AI systems greater autonomy.Memory PoisoningResearchers at US-based cyber security company Forcepoint have demonstrated how an attacker could manipulate an AI agent into storing malicious information in...