AI Security

AI Discovery Is Flooding Vulnerability Backlogs

Gartner’s 2Q26 survey put AI vulnerability discovery at the top emerging cyber risk, and Kaspersky said Q2 2026 brought an unprecedented surge in registered CVEs. Kaspersky also said new vulnerability classes appeared in Linux networking, and that OpenClaw was among the most affected AI projects, with more than 200 CVEs in the quarter.

The shift is speed and volume at once: AI tools can sift huge codebases for flaws and help turn them into working exploit code quickly, while the same tooling is also surfacing new bugs in AI products and AI-generated code. That shrinks the time between found and weaponized and pushes more issues into patch queues than normal review cycles can clear.

For teams running AI development stacks, Linux-heavy infrastructure, or AI products like OpenClaw, the exposure is now a backlog problem as much as a tooling problem. The reporting does not say the surge is slowing, only that discovery and exploitability are moving faster than the cadence most organizations use to absorb them.

4 sources · Aug 27

CVE-2026-25253

NVD KEV

CVSS 8.8 HIGH: openClaw (aka clawdbot or Moltbot) before 2026.1.29 obtains a gatewayUrl value from a query string and automatically… EPSS 24% (98th percentile).

Timeline

Sources

Part of the PlainSec briefing for 2026-08-26

Every edition of this story: AI Discovery Is Flooding Vulnerability Backlogs

More from today