What happens when an AI tool writes buggy code that ends up in Java? OpenJDK now has an answer, and it puts the responsibility squarely on the contributor.
The OpenJDK project - which maintains the open-source version of Java used in enterprise software, Android, and server-side applications worldwide - published an interim policy this week governing AI tool use in contributions. The rules are straightforward: you can use AI assistants to write or review code, but you must disclose when you did, and you own everything that comes from it. Bug, license violation, security flaw - it's yours, not the AI's.
The policy also flags intellectual property risk. AI coding assistants are trained on large amounts of code, some of it under licenses that restrict commercial or derivative use. OpenJDK's policy requires contributors to treat AI-generated code with the same scrutiny they'd apply to copying code from a random website - assume it might carry legal obligations until verified otherwise.
The "interim" label matters. This isn't a final position. OpenJDK leadership expects the policy to evolve as AI tools improve and as legal questions around AI-generated code get resolved in courts and legislation.
A Signal to the Rest of Open-Source
OpenJDK isn't the first open-source project to address AI contributions, but its reach makes this policy significant. Java runs on billions of devices and inside critical enterprise infrastructure. When a project of this scale formalizes AI rules, smaller projects will reference it.
The practical effect for contributors is modest right now: disclose AI use, review everything carefully, don't assume the output is clean. For project maintainers, it provides documented grounds to push back on low-quality AI-generated submissions - code that compiles but isn't well-reasoned or clearly understood by the person submitting it.