Research · curated 1 Oct 2026
Vulnerability Discovery and Exploitation Trends in the AI Era | Google Cloud Blog
First reported google.com
Coverage timeline
Single-source research — first reported, latest, and curated coincide.
Why it matters
GTIG's data shows AI is accelerating both the pace and consequence of vulnerability discovery and exploitation, including against the LLM/AI gateway stack, forcing defenders to shift from mass-patching to threat-intelligence-driven triage.
Google Threat Intelligence Group (GTIG) published a study measuring how AI is reshaping the vulnerability landscape, finding monthly CVE disclosures doubled from 5,045 (Jan 2026) to 10,740 (Aug 2026) and exploitation nearly doubled, while AI-assisted discovery surfaces proportionally more moderate-risk and RCE-leading vulnerabilities. The report also examines vulnerabilities targeting the AI/LLM operational stack (e.g., CVEs in LiteLLM, Langflow) and urges intelligence-driven triage with automated, agentic remediation.