Uploaded November 2025 | Updated September 2026, 1 week ago

---

Security researchers uncover new ChatGPT vulnerabilities that expose user data, Microsoft prepares “Agentic Users” with full M365 access, and the EU considers loosening GDPR rules around cookies and AI training. Today’s briefing breaks down the risks, policy shifts, and emerging challenges enterprises must prepare for as AI becomes deeply embedded in daily operations.

Read more on Computerworld and CSO Online:
Researchers trick ChatGPT into prompt injecting itself
csoonline.com/article/4086965/researchers-trick-chatgpt-into-prompt-injecting-itself.html

Microsoft product roadmap hints at ‘Agentic Users’ with their own M365 licenses
computerworld.com/article/4087533/microsoft-product-roadmap-hints-at-agentic-users-with-their-own-m365-licenses.html

European Commission moves to loosen GDPR for AI and cookie tracking
computerworld.com/article/4087347/european-commission-moves-to-loosen-gdpr-for-ai-and-cookie-tracking.html

Featuring host Arnold Davick:
linkedin.com/in/arnolddavick

Catch up on past episodes:
computerworld.com/videos/2-minute-tech-briefing

Follow TECH(talk) for the latest tech news and discussion!

------------------------------­----

Keith Shaw
linkedin.com/in/shawkeith

SUBSCRIBE: youtube.com/@CWTECHtalk

TIKTOK: tiktok.com/@todayintechpodcast