AI agents introduce a different identity problem because software can increasingly take actions instead of simply providing information. In #SmartManufacturing environments, that raises practical questions about what an agent can access, what it can change, and how those actions are recorded. Existing identity controls were largely designed around people and applications, so extending them to autonomous activity requires careful planning. It is best to establish clear boundaries and auditability early on rather than discover after the fact that an automated process has more access than anyone intended. #CyberProtection #IdentityGovernance #DataProtection #CIOInsights 

https://www.cio.com/article/4212025/your-identity-governance-wasnt-built-for-ai-agents.html