Over the past couple of days, I was digging into the latest Anthropic Threat Report and one section really hit me.

They wrote: ‘We’ve developed sophisticated safety and security measures to prevent misuse of our AI models. While generally effective, cybercriminals keep finding ways around them.’

And then they shared some eye-opening case studies—threat actors aren’t just asking AI for advice, they’re embedding it across their entire attack lifecycle. We’re talking reconnaissance, credential harvesting, extortion campaigns, even creating fake identities at scale. This is a whole new level of AI misuse—where a single actor can punch way above their weight class by turning AI into both consultant and operator.

That’s why I’m so excited about today’s guest: Clark Harshbarger, former Director of Incident Response at CrowdStrike. We’re going to explore both sides of this coin: how attackers are scaling their operations with AI, and how incident responders are starting to fight fire with fire—using AI to speed up detection and response when every second counts. Article: https://www-cdn.anthropic.com/b2a76c6f6992465c09a6f2fce282f6c0cea8c200.pdf