Oracle bans AI-generated code in OpenJDK contributions while championing it internally

Oracle has told OpenJDK contributors not to submit any material created by large language models, even as the company touts AI-written code for its own products.

Oracle issued a notice to the OpenJDK community forbidding any contributions that contain content produced, wholly or partially, by large language models or similar deep-learning systems. The guideline covers code, documentation, images and other artifacts in repositories, pull requests and issue trackers, citing potential burdens on reviewers and risks to safety, security and IP. While developers can employ AI tools for personal debugging, research or code comprehension, they must not push AI-generated output upstream.

This caution stands in stark contrast to statements from Oracle co-founder Larry Ellison, who said the company relies on AI models to write program logic, and co-CEO Mike Sicilia, who highlighted AI as a competitive advantage that shrinks engineering teams. Oracle also linked its AI push to a recent 21,000-job reduction, noting that automation may continue to trim staff. The move underscores the tension between open-source governance and Oracle’s aggressive internal AI adoption.