This week, Anthropic shipped its first public release of a large language model, Fable 5, based on its Mythos 5 frontier model. Mythos has been a hot topic of discussion in the tech industry, given it excels at finding software vulnerabilities – making it both a potentially useful resource for defenders and a potential accelerant for attackers. Todd Thiemann, principal analyst at Omdia, shares insights from his research into identity management and data security for AI and how enterprise agentic AI security should meet the Mythos moment.

Featuring: Todd Thiemann, principal analyst, Omdia

In today’s episode, we’ll cover…

• Mythos 5, Fable 5 and their cybersecurity implications
• Fighting AI attacks with AI defenses
• Balancing the security risks of using AI and of being vulnerable to AI
• How AI governance might factor into AI ROI

And more!

References:

• Cisco agentic AI security push faces enterprise trust gap
• Identity security for AI agents: The proliferation challenge
• Omdia Research Report: AI Agents and Identity Security: How Enterprises Are Rewriting the Rules
• Anthropic's Mythos raises the stakes for security validation

To learn more about software development and platform engineering, check out Search IT Operations.

To watch clips from our podcast, subscribe to our YouTube channel, @EyeOnTech.