What happened: A new major release of the Python character-encoding library chardet was described by its maintainer as a ground-up, AI-assisted rewrite under the permissive MIT license — triggering a dispute over whether the project can be relicensed away from its original LGPL terms.

Why it matters: If “AI clean-room rewrites” become a common path to relicense widely used libraries, open-source governance could shift from slow consensus to fast technical resets — with real consequences for who can reuse code, including in closed-source products.

Wider context: The legal status of AI-generated code is still unsettled, and the story exposes a second-order problem: even if a rewrite looks structurally different, model training data, developer exposure to prior code, and metadata carryover can all muddy what counts as a derivative work.

Background: The maintainer cited similarity checks showing low overlap with prior versions and described starting from an empty repository with instructions not to use GPL/LGPL code, while the original author argued that prior exposure and continuity of the project make the license change illegitimate.

AI can rewrite open source code—but can it rewrite the license, too? — Ars Technica

Singularity Soup Take: The uncomfortable truth is that “rewrite and relicense” is now cheap — and if the law doesn’t clarify what counts as taint in an AI-assisted workflow, open source will be governed less by licenses on paper and more by who controls the repo narrative.

Key Takeaways:

• Relicensing dispute: The maintainer described chardet 7.0 as a fresh MIT-licensed rewrite built with Claude Code, while the original author argued it must remain LGPL because it is effectively a continuation or derivative of the earlier project.
• Clean-room ambiguity: The maintainer acknowledged extensive exposure to the old codebase and cited low structural similarity metrics, but critics argue that LLM training data and any reused metadata/plans complicate claims that the result is truly independent.
• Community impact: The episode highlights how AI tools could enable fast rewrites of popular libraries — potentially changing the economics and trust model of open source, where licensing expectations historically travelled with the codebase over time.

Related News

AI-Assisted Rewrite Triggers Open-Source Licensing Fight — Singularity Soup’s earlier look at the same fault line between “new code” claims and license continuity.