Oracle bans AI-generated code from OpenJDK
Oracle, steward of the open-source OpenJDK project, has banned AI-generated code from OpenJDK contributions, citing safety, security and intellectual property risks. Under the new policy, developers may still use large language models privately to debug and review code, but they cannot submit AI-generated material to OpenJDK repositories, pull requests or other project channels.
The policy sits awkwardly next to Oracle's own public statements about AI use inside the company. Co-founder Larry Ellison has recently declared that AI models now write Oracle's code, and co-CEO Mike Sicilia has credited AI tools with letting smaller engineering teams deliver faster. So the company that says AI writes its internal software is simultaneously telling outside contributors they cannot submit AI-written code to one of its flagship open-source projects.
The ban lands against the backdrop of Oracle's heavy AI-driven spending: the company is investing $70 billion this year in datacentre expansion. That spending has drawn scrutiny from credit markets, and S&P downgraded Oracle's credit rating to BBB-, one notch above junk status, citing uncertain returns on the investment.
Key facts
- Oracle has banned AI-generated code from OpenJDK contributions, citing safety, security and intellectual property risks.
- Developers can still use LLMs privately to debug and review code, but cannot submit AI-generated material to OpenJDK repositories, pull requests or other project channels.
- The ban contrasts with co-founder Larry Ellison's recent claim that AI models now write Oracle's own code.
- Co-CEO Mike Sicilia has credited AI tools with letting smaller engineering teams deliver faster.
- Oracle is investing $70 billion this year in datacentre expansion; S&P downgraded its credit rating to BBB-, one notch above junk, citing uncertain returns on that investment.
Why it matters
OpenJDK is the reference implementation of Java, one of the most widely used programming languages in enterprise software. A formal ban on AI-generated contributions from its steward is one of the clearest public policy statements yet from a major vendor on where it draws the line for AI in open-source code, and it does so while that same vendor's leadership is publicly promoting AI as the engine behind its own internal development.
Who it affects
OpenJDK contributors and maintainers, who now face an explicit rule on what they can and cannot submit; and by extension the broader Java ecosystem and enterprises that depend on OpenJDK, since the policy governs the provenance of code entering a widely deployed runtime.
How to use it
Developers working on OpenJDK can keep using LLMs as a private aid, for debugging and for reviewing their own code. What they cannot do is submit AI-generated code itself into OpenJDK repositories, pull requests or other project channels; contributions have to be the developer's own work product.
How solid is it
The account comes from a single aggregated report citing The Register as the original source, and it is consistent internally: the stated rationale (safety, security, intellectual property risk) and the scope of the restriction (private use allowed, submission barred) are spelled out directly. What is missing is a precise date for when the policy took effect, the underlying policy document or announcement, and any on-record response from Ellison, Sicilia or Oracle to the apparent contradiction with their own public statements.
Risks and caveats
The source gives no figures on how many contributors or projects the ban actually touches, and no indication of how it will be enforced or detected in practice for AI-assisted versus AI-generated code, a line that can be hard to draw. The $70 billion datacentre figure and the S&P downgrade are separate facts about Oracle's finances; the source does not tie them causally to the OpenJDK policy itself.
“Oracle isn't writing”
— Larry Ellison, Oracle co-founder