AI-Assisted Research Drives Reported CVE Surge as Exploitation Timelines Shrink
Active Exploitation · 2 sources · Updated 1d ago · First seen Nov 16, 2026

Reported vulnerability disclosures exceeded 100,000 in 2026, with 100,222 CVEs published by November 13—more than double the 2025 total. Reporting citing VulnCheck attributes much of the increase to AI-assisted security research and automated flaw discovery. Approximately 1.4% of vulnerabilities are reportedly exploited in the wild, but the median disclosure-to-exploitation interval fell from 120 days in 2025 to 80 days in the first half of 2026. China-linked groups reportedly launched some attacks within 24 hours of disclosure, underscoring the much shorter response window for certain flaws.
The surge is increasing pressure on vendors and open-source maintainers to distinguish actionable findings from report volume. VulnCheck reportedly disclosed 14 critical vulnerabilities in Copeland refrigeration controllers, two of which were actively exploited. AI-generated vulnerability reports have also reportedly contributed to changes in curl’s reporting process and the closure of its bug bounty program. The reports describe CISA guidance favoring risk-based patch prioritization over severity scores or fixed remediation timelines alone, emphasizing internet exposure, evidence of active exploitation, and the potential to automate attacks. Security teams should use those factors to identify urgent remediation targets rather than treating every disclosure as equally pressing.


