Agentic Attackers, LLM Leaderboards, and Q-Day: A Conversation with F5 Labs
summary
F5 Labs publishes CASI and ARS leaderboards that score hundreds of models by running about 10,000 new attack prompts each month (now ~130,000+ prompts total) and by giving autonomous attackers goals under the agentic resistance score (ARS). Real-world incidents include an external researcher using agentic AI to harvest API keys and perform a blind SQL injection against an internal consulting chatbot. Research also flags slow post-quantum crypto adoption — about 8.5% PQC deployment across the top 1M sites — and argues Q-Day may arrive sooner than expected.
AI-written summary · about
related talks
34mThe dual role of AI in Cybersecurity
53mAI-equipped Threat Actors Versus AI-enhanced Cyber Tools: Who Wins?
57mAI as the New Hacker Frontier
01h 24mOffensive Cybersecurity by AI Promises and Pitfalls
39mThe AI Arms Race in Vulnerability Management, Who’s Winning?
20mCatch Attackers Before They Strike with AI-powered Threat Intelligence
40mAI As The New Hacker Frontier
51mThe Next Application Security Frontier: AI-Ready API Defense