Microsoft Tells Enterprises to Treat AI Agents Like Potential Insider Threats
Microsoft is urging organizations to stop assuming AI agents will always behave as intended, warning that agents with broad access to company systems can act unpredictably or be manipulated through hidden instructions. The company is expanding Project Perception, Microsoft Agent 365, and Entra Agent ID to give agents separate identities, tunable permissions, and dedicated "Guardian Agents" to police their activity. The post Microsoft Tells Enterprises to Treat AI Agents Like Potential Insider Threats appeared first on Secure Bulletin.
Itagora tiene in archivio titolo e sommario per la ricerca interna. Il testo completo resta sulla fonte. Il link in uscita non invia referrer.