AI Rush is Reviving Old Cybersecurity Mistakes, Mandiant VP Warns

**Executive summary:** A Mandiant executive warns that rapid enterprise AI adoption is reviving old security failures and creating new risks; red-team engagements revealed issues such as unencrypted AI–browser communication, ability to alter data classifications to bypass protections, and social-engineered initial access followed by AI-assisted exfiltration and policy changes. The report urges rapid implementation of AI security governance, secure architecture reviews, and red-team validation rather than assuming AI removes basic cybersecurity responsibilities.