More than half of AI-generated patches are broken
ID: 448bd243-1fc8-5455-a419-842ef8400095
STIX ID: report--448bd243-1fc8-5455-a419-842ef8400095
Feed Name: CyberScoop
Threat Score
Research from 1Password and Veracode found that current commercial large language models often fail to fully remediate high-impact CVEs and can introduce new vulnerabilities: 1Password reported a ~47% full-patch success rate for tested models and Veracode reported an average security pass rate around ~56%, with many models introducing OWASP Top 10 issues. The studies conclude AI-generated patches should be treated as code changes requiring human review and additional verification rather than trusted autonomous fixes.
Your team is not currently subscribed to this feed. You must subscribe to it in order to see this post.