Google Detects First AI-Generated Zero-Day Exploit Before Mass Attack

Google Threat Intelligence Group·high signal

Google's Threat Intelligence Group reported on May 11-12 that it identified and thwarted the first known case of cybercriminals using AI to discover and weaponize a zero-day vulnerability. The exploit targeted a Python-based open-source 2FA system; Google identified it by characteristic LLM artifacts including hallucinated CVSS scores and textbook-style docstrings. Bloomberg and CNBC confirmed the actors planned a 'mass exploitation event' but Google's proactive detection prevented deployment.