Cracking the Code: Unveiling Synergies Between Open Source Security and AI
summary
Session describes AIxCC, a DARPA and OpenSSF competition that requires teams to deliver open source tools which autonomously find and fix vulnerabilities in real projects (not synthetic code), scores prioritize patch effectiveness over mere discovery, and support multiple languages and vulnerability classes (memory corruption, logic, web flaws). Semifinals run at DEF CON with top semi teams winning $2M each; total prize pool is $30 million. Models bundled with tools must be open source; training data is not required.
AI-written summary · about
chapters (12)
- 0:00Open source security and AI
- 1:25DARPA and OpenSSF collaboration
- 5:45Open source security challenges
- 9:02Software security vulnerabilities
- 12:16Top-down vs bottom-up security
- 17:57Open source security strengths
- 19:45AI model security challenges
- 23:08AI for vulnerability discovery
- 27:12AIxCC challenge structure
- 36:11AIxCC scoring algorithm
- 42:27Q&A
- 51:25Call to action
related talks
44mArtificial Intelligence Cyber Challenge (AIxCC): Overview and Releasing Research as Open Source Software
11mAIxCC Results and New Open Source AI Projects To Help Secure Open Sou...
01h 02mPanel Discussion:The Present and Future of AI and Security
21mNavigating the Intersection: AI’s Role in Shaping the Secure Open Source Software Ecos...
17mLightning Talk: Navigating the Intersection: AI’s Role in Shaping the Secure Open Sour...
13mSecuring the Software Commons: Standards, Automation, and AI for a Resilie...
16mKeynote: Open Source for AI and AI for Open Source
05mIgnite: Security of Open-Source AI: is there any difference?
