CrowdStrike AI Security: Continuous Identity for Autonomous AI

CrowdStrike is launching a new capability called Continuous Identity for AI Agents. This system is designed to secure the autonomous enterprise by evaluating risk in real-time and eliminating standing privileges. Here's the thing: autonomous AI agents operate at superhuman speed. This creates unprecedented challenges for traditional security. CrowdStrike's CTO, Elia Zaitsev, states that "authorise once and trust indefinitely" is not a secure model. This new approach shifts from static access decisions to continuous identity. In other news, BT is the first UK company to join Anthropic’s Project Glasswing. This program gives critical infrastructure operators access to advanced AI tools. BT will use Anthropic's Claude Mythos Preview to strengthen its cyber defenses. The company currently blocks around four million cyber attacks daily. What's interesting is that Okta is also expanding its collaboration with Google Cloud. They aim to strengthen resilience across the modern workforce by combining their identity, cloud, and productivity solutions. Finally, Anthropic's Claude Fable 5, released on June 9th, was frozen by the US Government just three days later. This was due to fears about its automated hacking capabilities. Claude Mythos 5 is a non-public version of this technology, restricted for government use. The bottom line is that the rapid evolution of AI is fundamentally changing cybersecurity, requiring new approaches to identity and access management.