Home / Technology / AI Rewrites Code, Sparks Licensing Firestorm
AI Rewrites Code, Sparks Licensing Firestorm
11 Mar
Summary
- AI coding tools challenge traditional reverse engineering methods.
- A Python library rewrite using AI has ignited licensing disputes.
- The case questions AI's contribution to derivative works and copyright.
AI coding tools are introducing novel challenges to established practices like reverse engineering, particularly concerning "clean room" rewrites. The recent release of version 7.0 of the Python library chardet has brought these issues to the forefront. Maintainer Dan Blanchard described the update as a "ground-up, MIT-licensed rewrite" accomplished with AI assistance, boasting significant speed and accuracy improvements.
However, original author Mark Pilgrim disputes this, arguing that the new version is a derivative work and should retain the original LGPL license. Blanchard contends that while he had exposure to the old code, the AI generated code is structurally independent, citing low similarity statistics. He employed an "AI clean room" process, instructing the AI not to base its work on existing licensed code.
Complicating matters, the AI relied on metadata from previous versions, and its training data likely included prior chardet code. Blanchard's own extensive review and iteration on the AI-generated code further blur the lines. This situation highlights the unresolved legal questions surrounding AI authorship and copyright, with potential far-reaching implications for the economics and nature of software development.
