OpenJDK's Interim Policy on the Use of Generative AI

Oracle has approved an interim policy regulating the use of generative AI (LLMs, diffusion models, etc.) in the OpenJDK community's contribution process. The core of this policy is to prohibit directly contributing AI-generated content to OpenJDK repositories.

Key Prohibitions:

• Contribution of any content generated by LLMs or similar deep learning systems, including source code, text, and images, is prohibited
• Even if AI-generated code has been partially modified, it cannot be contributed if it still includes that code

Permitted Scope (Personal Use):

• Can be used as a personal learning and research tool for understanding, debugging, and reviewing code
• Use for analyzing existing code and research purposes is encouraged
• However, when sharing AI-generated output, it must be clearly labeled as AI-generated

Background for the Policy's Introduction:

• Reviewer burden: AI-generated "plausible-looking but incorrect code" excessively wastes reviewers' time
• Safety and security: To maintain the reliability and security of the JDK, which forms the foundation of the Java platform
• Intellectual property (IP): Copyright ownership of AI-generated output is unclear, and there is a risk of violating the IP rights requirements of the Oracle Contributor Agreement (OCA)