Summary
The question of AI attribution and provenance is unsettled but consequential — for licensing, for code review, and for institutional trust. Teams should adopt an explicit norm, even if the norm is 'we don't attribute.'
Objectives
- 01Explain the licensing risks of unattributed AI code.
- 02State the case for and against explicit attribution.
- 03Draft an attribution policy for a specific team.
The Lesson
The considerations
Some jurisdictions and licenses restrict the use of certain training-data-influenced outputs. Some enterprise customers require disclosure. Some open-source projects require CLA attestation. The norm to adopt is the one that survives your specific legal and cultural context.
Key Ideas
- Attribution is a policy question, not a technical one.
- Pick a norm and enforce it consistently.