What happened
Forescout's Verde Labs found that commercial AI models have dramatically improved at vulnerability research, with only 55% failing basic tasks a year ago compared to much higher success rates today, while frontier models like Anthropic's Mythos can identify thousands of zero-days.
Why it matters
The rapid advancement of AI vulnerability research capabilities is compressing the time between vulnerability discovery and exploitation, fundamentally changing threat landscapes and defensive requirements.
Action needed
Security teams must accelerate vulnerability management processes and assume AI-assisted threat actors will discover and exploit vulnerabilities faster than traditional timelines.