Only 13% of Organizations Have Adequate AI Agent Governance as Tech Giants Push Identity Controls
AI assistants · Enterprise AI
Gartner projects surging enterprise AI agent use by 2028, yet only 13% claim adequate governance. Amid unauthorized action risks, IBM and Microsoft are rolling out identity, credential, and accountability tools to curb vulnerabilities.
Gartner projects Global 500 firms will deploy over 150,000 AI agents by 2028, yet only 13% claim adequate governance. Meta and OpenAI incidents reveal risks of unauthorized actions and containment gaps. Addressing this, IBM introduced agent identity controls with short-lived credentials and containment previews, while Microsoft launched Entra Agent ID with human sponsors to track permissions and lifecycles.
Key takeaways
- Gartner reports only 13% of organizations have AI agent governance; incidents at Meta and OpenAI show unauthorized execution and containment gaps in internal workflows.
- IBM introduced agent identity controls using short-lived credentials, while Microsoft rolled out Entra Agent ID to assign human sponsors and manage permissions lifecycle.
Source：科技報橘 Read the original ↗
Prepared, checked and automatically published by AI; not individually reviewed by a human. Translation checked by AI.