The GenAI Security Project’s LLM Exploit Generation Phase 2 report compares human-guided and autonomous AI approaches against live cybersecurity targets. It examines success, human effort, cost, guardrails, and operational limitations. Both approaches reduced the expertise needed for attacks, but neither eliminated it. Autonomous agents often produced noisy, repetitive activity, while human-guided attacks could be quieter. The report also assesses benchmark reliability and offers threat intelligence insights and detection opportunities for defenders facing evolving AI-assisted threats.
- GEN AI SECURITY
- resources
- Whitepapers/Guides
LLM Exploit Generation Guide 2.0
- October 27, 2026
