Dean Ball, Anthropic's CISO, Patrick McKenzie, and others published warnings about AI-accelerated vulnerability surge and offense-defense imbalance
Record CVEs, PHP emergency, and NIST moves to modernize vulnerability database
AI is simultaneously the primary tool accelerating vulnerability discovery and a source of new vulnerability classes, while patching capacity has not kept pace and critical infrastructure systems often cannot be patched at all. The gap between discovery rates and remediation is drawing both industry warnings and a formal government response.
The full picture
The US National Vulnerability Database logged 45,207 flaws by late July 2026, nearly matching the record for all of 2025. Kaspersky's Q2 2026 analysis confirmed CVE registrations hit unprecedented levels that quarter, with AI tools playing a major role in discovering vulnerabilities in third-party software. Patch volumes have tracked the count: Google fixed 433 Chrome bugs in one July update versus 11 a year earlier, and Oracle patched 1,449 against 309. A PHP emergency tied to two AI-discovered flaws forced banks, airlines, and governments to take sites offline; Palo Alto Networks' Unit 42 estimates more than a third of 22 million internet-exposed systems vulnerable to those PHP flaws run software that cannot be patched, with upgrades expected to take months. Figures cited in October 2026 put 3% of all critical software security bugs in recorded history as found in the last month. Kaspersky also found that AI software itself introduces new vulnerability classes, with injection vulnerabilities, once mostly confined to web apps, now broadly relevant due to the rise of AI agents. NIST published a formal Request for Information on August 12, 2026, seeking public input on modernizing the National Vulnerability Database in light of AI, drawing 135 public submissions.
How it developed
Kaspersky's Securelist Q2 2026 report confirmed CVE registrations hit unprecedented levels, with AI tools driving vulnerability discovery and AI software itself introducing new vulnerability classes
NIST published a formal Request for Information on modernizing the National Vulnerability Database in light of AI, drawing 135 public submissions
Sources
Want this in your inbox?
I send one email each morning with the stories that moved. If you would rather just read here, that works too.
Subscribe free