Skip to content

AI_POLICY.md

Latest commit

 

History

History
35 lines (24 loc) · 2.32 KB

AI_POLICY.md

File metadata and controls

35 lines (24 loc) · 2.32 KB
Raw

AI Usage Policy

Responsible use of AI

Sync-in is a human-led open-source project that has been designed, developed, and maintained over several years. Its architecture, roadmap, and final technical decisions remain under the control of its maintainers.

We use AI in a controlled manner to support specific development tasks. It may help us understand and explore the codebase, draft tests, review changes, investigate issues, improve documentation, and assist with specific implementation tasks.

AI-generated output is not treated as authoritative and is never accepted blindly. Every accepted change must be understood, reviewed, adapted when necessary, and held to the same quality, testing, security, privacy, and licensing standards as any other contribution.

Because Sync-in handles user files, including those of its own maintainers, we take a cautious approach to AI-assisted code generation. Maintaining control over the codebase and limiting technical debt take precedence over development speed. AI may assist with implementation, but maintainers retain full responsibility for the code and its long-term evolution.

Secrets, credentials, personal data, private vulnerability reports, and other confidential information must never be shared with external AI services.

AI-assisted contributions

AI-assisted code generation is permitted, but it must remain strictly supervised and must not replace contributors' understanding or judgment.

When submitting an AI-assisted contribution, contributors must:

  • Understand and be able to explain every submitted change.
  • Review and, when necessary, adapt generated output before submission.
  • Run the relevant tests and checks and verify the expected behavior.
  • Follow the project's existing architecture, code quality, security, privacy, and licensing requirements.
  • Ensure that they have the right to use the AI tool and submit its output.
  • Identify any third-party material and its applicable license, as required by the Contributor License Agreement.

AI-assisted reviews may supplement, but never replace, human review. Suggestions or findings produced by an AI tool must be validated before they are acted upon. Maintainers may ask contributors to explain their changes, provide additional verification, or revise an AI-assisted contribution that does not meet the project's standards.