SANS Launches 'Find Evil!' Hackathon to Accelerate AI-Powered Cyber Defense
Event summary
- SANS Institute launched 'Find Evil!', a hackathon for autonomous AI incident response, with 1,100 participants registered.
- Rob T. Lee demonstrated AI-augmented forensic analysis in 14 minutes, a task typically taking weeks.
- Anthropic's Claude Mythos Preview discovered thousands of critical zero-day vulnerabilities, 99% unpatched.
- The two-month competition aims to make Protocol SIFT production-ready, with prizes exceeding $22,000.
- Submissions are due June 15, 2026, with winners announced around July 8, 2026.
The big picture
The 'Find Evil!' hackathon underscores the urgent need for AI-augmented cyber defenses as vulnerability discovery outpaces human remediation capacity. SANS is leveraging its 19-year-old community-built tooling to accelerate defensive capabilities, responding to state-sponsored AI-driven attacks and critical unpatched vulnerabilities. This initiative reflects a broader industry shift towards autonomous agents in cybersecurity, with potential implications for infrastructure protection and regulatory compliance.
What we're watching
- AI Defense Gap
- Whether SANS's community-driven approach can close the gap between AI-powered vulnerability discovery and defensive response.
- Hallucination Management
- The pace at which participants can develop robust hallucination management in autonomous AI agents.
- Industry Adoption
- How quickly the cybersecurity industry will integrate and deploy the solutions developed during the hackathon.
Related topics