CyberWire Daily

The email that tricked an AI.

14 snips
Sep 19, 2025
Karin Ophir Zimet, Chief People Officer at TORQ, discusses an innovative internship program aimed at enhancing AI skills and workforce literacy. She delves into training methods that teach coding with AI, fostering critical thinking, and improving speed and accuracy in product development. The conversation highlights the importance of preparing the workforce for an AI-driven future, alongside insights into the dynamic landscape of cybersecurity.
Ask episode
AI Snips
Chapters
Transcript
Episode notes
INSIGHT

AI Agents Can Be Tricked To Leak Data

  • AI agents that access user data can be manipulated by hidden instructions inside user content.
  • Radware's finding forced OpenAI to patch a ChatGPT agent that could exfiltrate Gmail data without user interaction.
ADVICE

Patch EPMM And Harden MDM Monitoring

  • Patch vulnerable MDM platforms immediately and monitor for abnormal activity.
  • Follow CISA guidance: strengthen MDM monitoring and adopt best security practices for EPMM.
ADVICE

Update Firebox Firmware Now

  • Update Firebox firewalls to fixed versions without delay.
  • Apply vendor patches promptly to avoid unauthenticated remote code execution risks.
Get the Snipd Podcast app to discover more snips from this episode
Get the app