I find it distasteful and disturbing that copyright infringement by the people training the LLM in violation of a license is considered contamination by the licensed code. It’s not contamination. The code didn’t seep into your codebase. If the LLM was trained in such a way that portions of code long enough to be protectable then the license was violated by humans. The liability for the problem doesn’t lie on the shoulders of the contributors to the originally licensed code. It lies on the people inserting it into your codebase without following the terms of the license.
The article also singles out the GPL repeatedly as a source of contamination. It doesn’t mention source-available proprietary licenses. It doesn’t mention code put online with no clear license, which according to the Bern Convention and the laws in at least the United States is automatically copyright protected with no license for use by others at all. It doesn’t talk about attribution for BSD-style or CC-SA-Attribution licenses. There’s no mention of leaked proprietary code. It just singles out GPL as some sort of unique problem.
This seems quite shoddy and biased for an article by someone who’s writing about the law.
The article also singles out the GPL repeatedly as a source of contamination. It doesn’t mention source-available proprietary licenses. It doesn’t mention code put online with no clear license, which according to the Bern Convention and the laws in at least the United States is automatically copyright protected with no license for use by others at all. It doesn’t talk about attribution for BSD-style or CC-SA-Attribution licenses. There’s no mention of leaked proprietary code. It just singles out GPL as some sort of unique problem.
This seems quite shoddy and biased for an article by someone who’s writing about the law.