Oracle Bans AI-Generated Code in OpenJDK Despite Internal LLM Use

Oracle has banned AI-generated code from OpenJDK contributions. The policy, reported by The Register and aggregated by Dealroom, explicitly prohibits submitting AI-generated material to repositories, pull requests, or other project channels. Developers can still use LLMs for private debugging and code review, but anything destined for the open-source Java project must be human-authored.

Policy Details

• Scope: Applies to all OpenJDK contributions, including code, documentation, and tests.
• Rationale: Oracle cites safety, security, and intellectual property (IP) risks associated with AI-generated content.
• Permitted use: LLMs can be used privately for debugging and reviewing code, but the output must not directly enter the project.

The Contradiction

The policy stands in stark contrast to Oracle's internal practices. Co-founder Larry Ellison has publicly declared that AI models now write Oracle's code. Co-CEO Mike Sicilia credits AI tools with enabling smaller engineering teams to deliver faster. This internal reliance on AI makes the OpenJDK ban feel inconsistent.

Oracle is not halting AI investments. The company plans to spend $70 billion this year on datacenter expansion to support AI workloads. That spending spree has caught the attention of credit agency S&P, which downgraded Oracle's rating to BBB- — one notch above junk status — citing uncertain returns on investment.

What It Means for Developers

If you contribute to OpenJDK, your pull requests must be free of AI-generated code. Using an LLM to help you understand a codebase or spot bugs is fine, but you can't paste LLM output directly into a PR. For everyone else, this highlights a growing tension between corporate AI adoption and open-source governance. Expect more projects to formalize similar policies.