| Mobile Web

Legacy security models cannot stop threats from AI spread; ambient and autonomous agents are the answer

As AI spreads inside companies, security risks are growing, and Microsoft is highlighting “ambient” and “autonomous” as key concepts. Mick Doone (믹 듄), a senior security adviser manager for Asia Pacific, said security must be built into systems from the design stage and operate continuously across identity, endpoints, cloud, data and AI agents. He said prompt and response tampering and agent-connected plugins can become attack paths.