Anthropic's Claude Mythos Triggers Surge in Critical Vulnerabilities

Research from Check Point shows that AI‑augmented threat actors and automated scanning tools have doubled the rate of critical CVE discoveries, creating an "AI‑Driven Vulnerability Inflation" that overwhelms security operations centers with low‑priority alerts. Only about 8.3% of reported critical vulnerabilities now require immediate remediation, widening an exposure gap between high‑level AI security policies and frontline defenses.

The spike follows Anthropic’s April 2026 announcement of the Claude Mythos Preview, which can autonomously discover software flaws. In June, disclosed high‑ and critical‑severity CVEs rose more than 3.5 times the previous monthly record, with Project Glasswing partners—including Microsoft, Google, Apple and AWS—reporting over 10,000 such vulnerabilities found using the model. Similar initiatives are underway at OpenAI. The rapid increase raises concerns about alert fatigue and the need for risk‑based prioritization in vulnerability management.