Anthropic Claude Code Leak Tests AI Copyright Issues
Sigrid Jin, a 25-year-old undergraduate student at the University of British Columbia in Vancouver, Canada, spotted news about a major leak from Anthropic while waiting to board a plane. The artificial intelligence company had accidentally exposed the source code for Claude Code, its well-known AI agent. Mr. Jin quickly posted a copy of the code online. His girlfriend sent a concerned text, asking if he was breaking copyright law.
To address the issue, Mr. Jin instructed a group of AI assistants to convert the leaked code into a different programming language. He then uploaded this altered version to the web. In just hours, over 100,000 people liked or shared links to it.
The Leak's Origin and Anthropic's Response
Anthropic ranks among top AI firms, similar to OpenAI. The company attributed the leak to a human mistake. It contacted GitHub, the popular online repository for computer code, and requested removal of posts containing the code, pointing to copyright breaches. GitHub complied by deleting thousands of such posts.
Mr. Jin's rewritten code, however, stays available. He noted that Anthropic has not contacted him to request its removal. Anthropic did not answer questions from The New York Times on the matter.
Anthropic, founded in 2021 by former OpenAI staff members, emphasizes AI safety and has received major funding from companies like Amazon. Its Claude series of models powers various tools, including the Claude Code agent designed for coding tasks. Such incidents raise questions about handling sensitive code in fast-growing AI development.
Rewriting Code with AI Assistance
Stay ahead of the AI curve
The most important updates, news, and content — delivered weekly.
No spam. Unsubscribe anytime.
Mr. Jin viewed the rewrite as creating an entirely new piece of work outside Anthropic's ownership. Reporters reached him from Taipei, Taiwan, where journalist Meaghan Tobin covered the story on April 22, 2026.
His actions stemmed not from a pursuit of profit or recognition but from a wish to spark discussion on a key issue. In an age where AI replicates computer code, artwork, music, and books in moments, what role does copyrighted material play?
Philosophical Questions on Creative Property
Mr. Jin aimed to highlight ethical concerns tied to AI agents. "I just wanted to raise some ethical questions in the A.I. agent era," he said. "Any creative work can be reproduced in a second."
It remains uncertain if Anthropic sees a difference between the original leaked code and Mr. Jin's version. The case tests boundaries of copyright law amid rapid advances in AI tools that speed up copying creative content.
GitHub, owned by Microsoft since 2018, serves millions of developers worldwide as a hub for storing and sharing code. Leaks like this one underscore vulnerabilities in code management, especially for proprietary AI projects.
The episode reflects ongoing debates in the AI field. Companies invest heavily in developing models like Claude, yet tools built on similar technology now challenge traditional protections for intellectual property.
Related on Neura Market