Safety · The Decoder ·

AI finds plenty of security flaws, but almost none of them get exploited

AI finds plenty of security flaws, but almost none of them get exploited

VulnCheck found that 14 of 1,061 vulnerabilities identified by AI in the first half of 2026 were linked to confirmed attacks, a 1.3% rate matching vulnerabilities overall. The median time to exploitation fell from 120 days to 80 days.

Read the full story at The Decoder →