Relicensing with AI-Assisted Rewrite (tuananh.net)

Article summary

The maintainers of the chardet project used an AI tool to rewrite the codebase and relicense it from LGPL to MIT, sparking a debate about the legitimacy of this approach. The original author claims this is a GPL violation, as the AI was trained on the original LGPL code. The US Supreme Court's recent decision on AI-generated material has added to the complexity of the issue. This case raises questions about the future of open-source licensing and the role of AI in software development.

Main themes

• AI-assisted relicensing
• Open-source licensing
• GPL and LGPL
• Clean room implementation
• AI-generated code
• Copyright law

What commenters say

• The use of AI to rewrite code and change licenses may be a way to circumvent GPL requirements, potentially threatening the concept of Copyleft.
• The 'clean room' process, which requires two separate teams to ensure no direct copying, may not be applicable when AI is used to generate code.
• Some argue that AI-generated code should be considered public domain, as it is not created by a human author, while others claim that the original license still applies.
• The Supreme Court's decision on AI-generated material has created a legal paradox, as AI-generated code may not be eligible for copyright, rendering licenses moot.
• The line between 'public data in' and 'public domain out' is blurred, and it is unclear whether AI-generated output should be considered public domain or still bound by the original license.
• The use of AI to generate code raises questions about authorship and ownership, and it is unclear who should be considered the author of AI-generated code.
• There are concerns that allowing AI-assisted relicensing could lead to a loss of rights for creators and developers, and potentially undermine the open-source community.
• The issue of AI-generated code and licensing is complex and multifaceted, and will likely require new laws and regulations to clarify the situation.