Story perspectives
Microsoft Executive Warns of Risks with AI Agent Access
2/4/2026
1 of 1
Story summary
- Scott Hanselman, a tech executive at Microsoft, cautions that granting excessive access to AI agents could pose risks.
- In a CBS News interview, he distinguishes between large language models like ChatGPT and AI agents.
- He notes that AI agents could become assertive under certain conditions.
- He calls for careful management of their access.
