Story perspectives

Experts Warn of Security Risks in Microsoft Copilot AI

3/23/2026

Summary

Story summary

Microsoft's Copilot AI has drawn scrutiny after errors, including hallucinating police reports and exposing sensitive information.
Dennis Xu of Gartner argued for banning Copilot on Friday afternoons, saying workers may overlook its mistakes.
Xu identified five security risks at a Gartner panel in Sydney, focusing on Copilot's tendency to compromise sensitive data.
The discussion framed these risks as intensified by AI as Copilot's use expands in corporate and government environments.

1 of 1

Story summary

Microsoft's Copilot AI has drawn scrutiny after errors, including hallucinating police reports and exposing sensitive information.
Dennis Xu of Gartner argued for banning Copilot on Friday afternoons, saying workers may overlook its mistakes.
Xu identified five security risks at a Gartner panel in Sydney, focusing on Copilot's tendency to compromise sensitive data.
The discussion framed these risks as intensified by AI as Copilot's use expands in corporate and government environments.