Cybersecurity
OpenAI Launches Patch the Planet to Secure Open-Source Code as AI Bug-Hunting Accelerates
The initiative pairs AI security tools with human expertise to help resource-strapped open-source maintainers defend against AI-powered vulnerability discovery.
WIRED — AI
OpenAI unveiled a comprehensive effort called Patch the Planet on Monday, partnering with security firm Trail of Bits and vulnerability management platforms to address a growing crisis: AI-powered vulnerability hunters are outpacing the capacity of volunteer-run open-source projects to respond.
The problem is real. Open-source maintainers, already stretched thin, are drowning in bug reports—many generated by AI tools and of questionable quality. Patch the Planet aims to flip that dynamic by providing free security consulting, code assessments, and patch development directly to maintainers.
The Initiative's Scale#
More than 30 projects are already enrolled, with Trail of Bits dedicating roughly a fifth of its workforce to the effort. In the opening week alone, the team uncovered hundreds of bugs and shipped dozens of patches. The program offers six months of free ChatGPT Pro access, subsidized Codex Security scanning, and tailored AI agents that maintainers can deploy long-term.
OpenAI's cyber tech lead Fouad Matin frames the work as damage control: "what we've effectively done is make it as efficient from a token perspective as possible to reduce the burden for maintainers." The company is underwriting 20 trillion tokens of security scanning for both open-source and private codebases.
The Competitive Context#
The announcement lands as OpenAI competes with Anthropic, whose Mythos 5 and Fable 5 models were pulled from the market this month under White House pressure over cybersecurity export concerns. OpenAI's latest GPT-5.5-Cyber checkpoint scores 85.6% on the CyberGym benchmark—beating Anthropic's 83.8%—though it remains in limited access only.
The Five Eyes intelligence alliance issued an unusually urgent warning Monday: frontier AI models will soon "exceed current industry expectations" in both offensive and defensive capabilities. "The timeline is not years, it is months," they cautioned.
Patch the Planet represents OpenAI's bet that the answer isn't restricting AI security tools—it's democratizing them.
