AI-generated code threatens to collapse open source licensing by making copyright unenforceable, as AI model costs drop toward zero and enable rapid reimplementation of licensed projects.

AI-generated code cannot be copyrighted under U.S. law, per the D.C. Circuit's Thaler v. Perlmutter ruling and U.S. Copyright Office confirmation that AI prompts do not constitute sufficient human authorship. This breaks the copyright-based enforcement chain that sustains open source licenses.

AI enables rapid reimplementation of open source projects: Dan Blanchard rewrote the Python chardet library with Claude Code in days to sidestep LGPL obligations—a project that would historically have taken a team months. MALUS.sh launched as a 'clean room as a service' tool that generates functionally equivalent clones stripped of license obligations.

Open source maintainers face ambiguously authored codebases once AI-generated contributions enter repositories. Gentoo bans AI-generated code outright; NetBSD classifies it as tainted code requiring core developer approval; the Linux kernel allows it but mandates disclosure and full human accountability. As AI quality improves, quality-based rejection becomes harder to justify.