What legal worries come up with AI-written code, and how do teams respond?

Summary of public sourcesWritten by our research agent from 5 public sources. Not first-hand experience.

Three legal worries come up. Pure AI output may not be copyrightable. Generated code may copy licensed code. And compliance teams ask where code and data go. The U.S. Copyright Office's 2025 report says copyright does not cover purely AI-generated material. It says prompts alone do not give enough control. Human selection, arrangement and edits can still be protected, case by case. NIST's generative AI profile lists intellectual property among its main risks.

The responses are practical. Document the human design, review and edits. Practitioners say trade secret protection matters more than copyright for most code. On copying, GitHub offers a filter that blocks suggestions matching public code. It says its indemnity covers customers who use that filter. Its code referencing feature instead shows matches and licenses, and the indemnity extends to it if teams comply with the cited licenses. On compliance, practitioners say the real controls are who can reach regulated data and how code reaches production.

Still open: what a vendor's legal cover changes in practice. No source here described a legal team's sign-off process.

Sources

• Hacker News: Ask HN: Is all AI coded code in the public domain?
• Hacker News: Ask HN: How do companies that use Cursor handle compliance?
• Standard: Copyright and Artificial Intelligence, Part 2: Copyrightability
• Vendor docs: Code referencing now generally available in GitHub Copilot and with Microsoft Azure AI
• Standard: Artificial Intelligence Risk Management Framework: Generative Artificial Intelligence Profile (NIST AI 600-1)