AI finds bugs faster than humans can fix them: how enterprise security teams must adapt
Google/DeepMind
Apple
Microsoft
Chainguard
DeepSeek
AI-driven vulnerability discovery is outpacing remediation, overwhelming security teams with a flood of reports. Google, Apple, Microsoft, and Linux maintainers are all struggling to keep up, and the volume of patches is rising, creating a triage crisis that threatens enterprise security.
The article highlights a growing imbalance: AI is finding security bugs at a pace far exceeding humans' ability to patch them. This is evidenced by Google fixing more Chrome bugs in June 2026 than in the past two years, and Apple restricting the number of bug submissions from researchers. Microsoft's July 2026 Patch Tuesday shipped 570 patches, including three zero-days, setting a record. The root cause is that AI makes it cheap to find flaws, but fixing them remains difficult, as a study shows LLMs introduce nearly 9x more new vulnerabilities. Security teams face a triage tax, spending time validating reports rather than fixing issues. Companies like Google suggest narrowing AI's scope, separating remediation from verification, and keeping humans in charge of complex changes. The article warns that the patching pace will only accelerate, and organizations must adopt stronger automation for deduplication and scoring to cope. It cites experts like Chainguard's CEO, Linux maintainer Greg Kroah-Hartman, Linux Foundation CEO Jim Zemlin, and Google's former Project Zero manager Ben Hawkes. It also mentions an attack using the DeepSeek LLM via the Hermes Agent framework. The article concludes that companies need to invest more in security despite budget constraints being the top cause of staffing shortages.
- Abbreviations
- CVE = Common Vulnerabilities and Exposures
- LLM = Large Language Model — большая языковая модель
- CISO = Chief Information Security Officer — директор по информационной безопасности
Source: ZDNet AI —
original