Google Confirms AI-Powered Zero-Day Discovery Bypassing 2FA for First Time

According to Decrypt, Google’s threat intelligence team recently confirmed that cybercriminals used artificial intelligence to discover and exploit a zero-day vulnerability in a popular open-source web administration tool that bypasses two-factor authentication. This marks the first time Google has confirmed AI-assisted zero-day vulnerability development in real-world attacks.

The vulnerability exploits a logic error rather than a code flaw. Unlike traditional scanners that search for crashes or malformed code, the AI analyzed the software’s intended behavior and detected logical contradictions, allowing attackers to bypass security checks without breaking encryption. Google coordinated with the affected vendor to patch the vulnerability before large-scale exploitation could occur.