Winners of DARPA’s AI Cyber Challenge
Finalists from DARPA's AI Cyber Challenge present autonomous cyber reasoning systems designed to find, patch, and remediate critical software vulnerabilities.
Summary
The DARPA AI Cyber Challenge (AIxCC)
DARPA launched the AI Cyber Challenge to spur the development of next-generation Cyber Reasoning Systems (CRS) that combine artificial intelligence with automated program analysis to secure critical open-source software.
Autonomous Vulnerability Discovery and Patching
Competing teams showcase CRS architectures that automatically analyze complex C, C++, and Java codebases, discover zero-day vulnerabilities, and generate verified, non-breaking security patches in real time.
Integrating LLMs with Dynamic Analysis Tools
The winning systems integrate large language models with fuzzing engines (such as AFL++ and libFuzzer), symbolic execution frameworks, and automated regression testbeds to eliminate false positives.
The Future of Automated Software Defense
Automating vulnerability discovery and patch deployment transforms software defense, enabling rapid remediation of supply chain vulnerabilities before adversaries can weaponize them.
Related talks
Main Stage·21 min
Perri Adams
DARPA Program Manager Perri Adams and industry leaders announce the AI Cyber Challenge (AIxCC) to automate software vulnerability discovery and remediation.
Creator Stage·22 min
D. Jurado · J. Nogue
Chronicles the deployment of XBOW, an autonomous AI offensive agent that discovered over 200 open-source vulnerabilities and submitted 1,000 bug bounty reports.
Previews·3 min
Andrew Carney
Andrew Carney previews the semifinals of the DARPA AI Cyber Challenge (AIxCC), pitting autonomous AI systems against software vulnerabilities.