FFmpeg Developer Accuses OxideAV of AI License Laundering in MagicYUV Issue

An FFmpeg developer (richardpl) has publicly called out OxideAV on GitHub for what appears to be an attempt to launder code licenses using AI. The issue, filed in the oxideav-magicyuv repository, questions the missing reverse-engineering documentation and the overall legitimacy of the project's license.

The Core Complaint

The developer asks: "Where is docs/video/magicyuv/magicyuv-trace-reverse-engineering.md located?" This file is referenced somewhere in the repo's documentation or code but is missing from the repository. The implication is that OxideAV may have used AI to reimplement the MagicYUV codec without properly documenting the reverse engineering process, which is required for GPL compliance when deriving from GPL-licensed FFmpeg code.

License Laundering with AI

License laundering is the practice of taking GPL-licensed code, passing it through a tool (like an AI model), and releasing the result under a different license — often a permissive one like MIT or Apache. The output may no longer contain literal copies of the original code, but the derived nature still legally requires GPL compliance. The FFmpeg developer suspects OxideAV used AI to regenerate the codec implementation, circumventing the license notice and attribution requirements.

What's at Stake

If the accusation holds, OxideAV could be legally liable for copyright infringement. The issue remains open with no response from the repository maintainers as of this writing. The HN community has flagged this as a growing pattern where AI-generated code is used to bypass open-source licenses.

This is not just a single bad actor — it underscores a systemic risk: as AI coding agents become more capable, we will see more attempts to launder code through them. Developers using AI tools should verify the provenance of training data and ensure their outputs remain license-compliant.