Microsoft Defender for Office 365: Prompt injection protection for email

🚨 The Signal: Microsoft Defender for Office 365 now detects and blocks malicious prompt injection attempts in emails. This protects AI assistants and agents from manipulation, preventing data exfiltration and system prompt exposure, enhancing enterprise AI security.

The Impact

Organisations with Defender for Office 365 Plan 2 or M365 E5 are affected, gaining automatic protection against AI prompt injection attacks via email.

• Security Teams: Reduced risk of AI agent manipulation and data exfiltration.
• AI Users: Enhanced trust in AI assistant interactions, less risk of compromised outputs.
• Admins: Automatic quarantine of prompt injection emails, reducing manual intervention.
• Organisations: Improved compliance with data protection and AI governance policies.

Domain: Agentic-AI · Impact: high · Workload: Microsoft Defender