WHEN THE CODE GETS CREATIVE: ANTHROPIC, AI OWNERSHIP, AND THE LAW

 By Jessica Debrah

In April 2023, a song called Heart on My Sleeve shot across TikTok and streaming platforms like wildfire. It sounded like Drake and The Weeknd teaming up for a surprise drop. Except neither artist had ever stepped into a studio for it. The track was written, composed, and “performed” entirely by AI.

A few years earlier, in Christie’s auction room, a solemn-looking gentleman with blurry edges portrait titled Edmond de Belamy sold for over $432,000 and yet again the “artist” was not human, but a generative algorithm trained on thousands of other works.

From viral songs to record-breaking paintings, AI is no longer just assisting creativity, it’s producing it. Who then owns these works? Who gets the royalties?
And in a world where the “creator” could be a line of code, does the current Copyright Laws even have the right tools to keep up?

Anthropic isn’t a household name like ChatGPT or Midjourney, but in the AI world, it’s one of the power houses pushing the frontier of generative intelligence. Founded in 2021 by former OpenAI researchers, it is an AI research and development company which has developed a suite of large language models named Claude as a competitor to OpenAI's ChatGPT. The company focuses on safety and ethics, encouraging responsible innovation across the artificial intelligence industry. This generative AI is capable of generating essays, poetry, product designs, and even business strategies with a level of polish that humans express.

The U.S. Copyright Office made its stance clear that works created entirely by AI without “meaningful human authorship” are not eligible for copyright. Other jurisdictions, like the UK, have taken a more flexible approach, granting rights to the human who “made the arrangements” for the AI’s creation. Ghana and other African IP regimes are still evolving in this space, meaning the rules are even less defined.

Copyright laws exist to protect human creativity. But AI isn’t human. It doesn’t sleep, doesn’t need inspiration, and can create endlessly. So if an AI like Claude produces a poem, a design, or a soundtrack, the law runs into a wall. Can the person who typed the prompts claim full rights? Or should there be a new category of ownership entirely? The answers to these questions will shape the creative economy for decades.

For Anthropic, they are faced with several lawsuits and policy debates about whether its training data, often copied from existing works could infringe on the same copyrights that IP law is supposed to protect.  Meanwhile, its models are empowering creators and businesses to generate valuable assets at unprecedented speed and scale.

The Legal Landscape: Anthropic vs. Copyright

 In June 2025, U.S. District Judge William Alsup delivered a landmark ruling in the Bartz v. Anthropic case. He ruled that using legally purchased books to train Claude’s AI models qualifies as fair use, calling it “exceedingly transformative” and likening it to how a writer learns from reading. Nonetheless, Judge Alsup drew a firm line when it came to piracy, opposing Anthropic’s act of downloading over 7 million pirated books from sites like LibGen. It was said that it isn't protected under fair use, moving that part of the case ahead to trial.

This is a pivotal moment for AI and IP. When AI like Anthropic’s Claude learns from millions of works, it isn’t “copying” in the classic sense but can be understood as remixing patterns into something new. The fair use ruling says, “Okay, that’s legal if you actually paid for the stuff you learned from.”

But the moment the issue of piracy sets in, the innovation story becomes theft. And that’s where the tension sits: tech companies want room to experiment, while creators want the law to stop their life’s work from becoming free training data for someone else’s idea and benefit.

Looking Ahead

AI isn’t slowing down; neither are the legal battles. What we’re watching is the early blueprint for how the world will treat creativity in the machine age. Take Thomson Reuters v. Ross Intelligence, where Reuters claims its legal database was unfairly used to train an AI tool without permission. Or Getty Images V. Stability AI, where the stock photo giant argues its copyrighted images were scraped to generate AI art. Together, these cases are setting the stage for how far AI can go when it borrows from human creations.

The next few years will decide whether the law can keep pace with code that learns, creates, and competes with humans. For creators, it’s a call to register your rights, know your worth, and push for policies that protect your ideas.

Because in this new era, the question isn’t just “Can AI create?” Because it can create endlessly. The question rather becomes “Who gets to own what it creates?”

And the answer will shape the future of art, tech, and every bright idea yet to be imagined.