AI Finds 1,000+ Flaws but Only 1.3% Get Exploited, VulnCheck Reports

The Decoder · rss · 2026-08-02

Data from security firm VulnCheck shows that while AI is active in discovering security flaws, the actual exploitation rate of these vulnerabilities is extremely low. Out of 1,061 AI-discovered vulnerabilities in the first half of 2026, only 14 saw confirmed attacks—about 1.3%, roughly matching the exploitation rate of vulnerabilities overall.

However, the report notes that the speed of exploitation is accelerating, with the median time from discovery to exploit dropping from 120 days to 80 days.

Original post →

AI Finds 1,000+ Flaws but Only 1.3% Get Exploited, VulnCheck Reports

More from Safety