Anthropic's 512,000-Line Claude Code Leak Rekindles AI Code Copyright Fight as Copilot Appeal Advances
Updated
Updated · O'Reilly Media · Jun 15
Anthropic's 512,000-Line Claude Code Leak Rekindles AI Code Copyright Fight as Copilot Appeal Advances
2 articles · Updated · O'Reilly Media · Jun 15
Summary
512,000 leaked lines of Claude Code turned a niche legal issue into an immediate test of whether predominantly AI-written code can be copyrighted or support DMCA takedowns.
U.S. doctrine still ties copyright to human authorship: Thaler left fully AI-made works unprotected, but no court has yet squarely ruled on AI-generated code or how much human editing makes it protectable.
That uncertainty extends beyond ownership because employers may still claim AI-assisted work under broad IP-assignment clauses, including some side projects built with company-licensed tools.
Open-source risk is a second front: if an AI tool reproduces substantial GPL-licensed code verbatim, developers could trigger copyleft obligations even without knowing the source.
Doe v. GitHub, now in the Ninth Circuit after most claims were dismissed below, has already pushed code-license scans and output filters into standard commercial and M&A diligence.
If AI writes your code, do you truly own your company's most valuable asset?
When AI is your co-author, what human input is needed to legally own your creation?
Anthropic’s $2.5B Claude Code Leak: Security Failures, Legal Battles, and the Future of AI IP
Overview
On March 31, 2026, Anthropic accidentally leaked its proprietary 'Claude Code' due to a missed manual step in the deployment process, which was quickly traced to human error rather than AI involvement. The staffer responsible was not fired, as the company recognized it as an honest mistake. While the leak did not expose the core model weights, it allowed technically skilled individuals to access sensitive internal information, including development roadmaps and system architecture. This incident highlighted vulnerabilities in Anthropic’s software supply chain and sparked debates about intellectual property, legal risks, and the challenges of protecting AI-generated code in a rapidly evolving industry.