Anthropic DMCA Notice on GitHub Accidentally Removes Legitimate Code Forks

Anthropic, the artificial intelligence company behind the Claude language model, acknowledged that a recent effort to remove leaked source code from GitHub unintentionally swept up many legitimate forks of its own public repository. The incident has highlighted the challenges technology firms face in balancing copyright enforcement with open-source collaboration.

Earlier this week, Anthropic submitted a Digital Millennium Copyright Act (DMCA) notice to GitHub requesting the removal of repositories containing unauthorized copies of its recently leaked Claude Code client source code. The leak, previously reported, exposed key elements of code behind one of Anthropic’s main AI services.

The DMCA request identified nearly 100 forked repositories directly linked to the original leak, posted by a GitHub user. However, GitHub acted on the notice by taking down a much broader network of over 8,100 forks. According to GitHub, the removal extended to this wider group because "the submitter alleged that all or most of the forks were infringing to the same extent as the parent repository."

Many of the repositories removed did not contain any of the leaked code. Instead, they were legitimate forks of Anthropic's official public Claude Code repository. This codebase is openly shared by the company to promote transparency and to encourage outside contributions, such as bug reports and suggested fixes.

The expanded removal drew complaints from developers and open-source contributors. Several affected users voiced their frustration on social media, stating they were wrongly caught in the copyright dragnet despite not distributing any of the leaked or proprietary code.

Anthropic has since reversed the overbroad DMCA takedown, reinstating repositories that did not host infringing material. However, the episode underscores how copyright enforcement actions can risk collateral damage within software development communities, particularly in the context of open-source AI projects. Such incidents emphasize the sensitivity around intellectual property management as AI development becomes increasingly collaborative and decentralized.

Although the immediate issue has been resolved, Anthropic continues to grapple with the wider challenge of containing the unauthorized distribution of its leaked code. The case draws attention to the need for clear protocols and careful coordination between AI companies, online platforms, and open-source communities to safeguard intellectual property without stifling legitimate contributions.

Source: arstechnica.com