[AI Hacking in M365] Defender Stops Prompt Injection & Jail Breaking

Опубликовано: 27 Июнь 2026
на канале: Threatscape
1,126
41

🔍 Purview Implementation Service: https://www.threatscape.com/threatsca...

🔍 Copilot for M365 Readiness Assessment: https://www.threatscape.com/copilot-f...

AI jailbreaks and prompt injections are no longer hypothetical — they’re happening.

In this video, Roo Campbell (Microsoft MVP) walks you through:

A real jailbreak attempt on Microsoft 365 Copilot

How prompt injection works through Outlook emails

Defender XDR’s new detections (with KQL walkthrough)

How Microsoft Purview helps trace full Copilot sessions

Practical hunting tips for AI threats in your environment

This is a must-watch for security professionals defending AI-integrated SaaS tools. Learn how to get ahead of prompt-based attacks with Defender and Purview.

🔍 Complimentary Advisory Services [no-obligation consultation]:

Microsoft Purview Advisory Service: https://www.threatscape.com/microsoft...

Microsoft Defender XDR Advisory Service:
https://www.threatscape.com/microsoft...

📩 Like and subscribe for weekly threat detection deep-dives.

Website: https://www.threatscape.com/
Linkedin:   / 942506  
Email Address : [email protected]