
Read the High row again before believing the AI claim
2026-11-21AI
Google has published the first severity breakdown of AI-discovered vulnerabilities against human-found ones. Everybody will read it as proof that models find worse bugs. The High-severity row is three percent against four.