Anthropic says it found 5,500 verified vulnerabilities in April-October and Glasswing partners found 129K+ in April-July; 33K+ were critical or high severity (Reuters)

Anthropic’s report of thousands of verified vulnerabilities, along with a much larger set found by its partner program, underscores how quickly AI systems are becoming a meaningful security risk surface for enterprises. For CIOs and technology leaders, the strategic implication is clear: adopting powerful AI models now requires the same discipline as any other critical platform, including continuous red-teaming, strict governance, and tighter controls over how models are tested, deployed, and monitored. IT organizations should expect AI security to become a standing operational function rather than a one-time review.