AI & Enterprise
Experts say poor security management behind AI agent hacking incidents
Experts said hacking incidents involving AI agents from OpenAI, Anthropic and Google point to poor security management by AI companies, not an AI “runaway” problem. Japanese outlet Gigazine reported that CSO Online reviewed causes and accountability based on analyses by six security experts. They cited weaknesses in sandbox design, isolation and monitoring. Experts urged tighter access controls, network segmentation and real-time monitoring using proven security practices.
