Landmark Court Decision Lets AI Train on Copyrighted Books, Leaving Authors Furious

Landmark Ruling in Favor of AI Raises Concerns for Creatives

A recent court decision has set a precedent that could deeply affect authors and other creatives. U.S. District Judge William Alsup ruled that Anthropic’s use of millions of copyrighted books to train its AI falls under “fair use.” This is the first ruling of its kind and opens the door for large language models (LLMs) to be trained on copyrighted materials, provided those works were legally acquired.

Judge Alsup described the AI technology as “among the most transformative many of us will see in our lifetimes.” He noted that the AI models do not replicate authors’ exact styles or specific books but use them to generate entirely new text. He emphasized that training AI on copyrighted works is “quintessentially transformative,” much like a reader learning to write by reading.

What This Means for Creatives

For authors like Andrea Bartz, Charles Graeber, and Kirk Wallace Johnson—who sued Anthropic—the ruling is a blow. They argued that Anthropic would profit off millions of copyrighted works without fair compensation. Anthropic responded by stating their AI models are designed not to copy or replace original works but to create something new.

According to intellectual property lawyer Daniel Barsky, this decision will likely be cited by other AI developers to justify using large copyrighted datasets for training foundational models. However, the story doesn’t end here.

Anthropic’s Legal Trouble Continues

Despite the fair use ruling, Anthropic faces another lawsuit for having initially downloaded 7 million copyrighted books illegally before purchasing them. This case is set for trial in December and could result in massive damages. Copyright infringement penalties can reach up to $150,000 per book, which means potential liabilities could reach unprecedented amounts.

Meta’s Case Adds to the Debate

In a separate case, authors sued Meta for allegedly using pirated books to train its AI. Despite clear evidence of unauthorized use, the court ruled in Meta’s favor. Plaintiffs’ attorneys expressed strong disagreement with this outcome, highlighting ongoing tensions between content creators and AI companies.

• AI companies can currently use copyrighted works for training with limited legal consequences.
• However, the legal landscape remains uncertain and could change if higher courts intervene.

For creatives, this means vigilance is necessary. The ability of AI to learn from copyrighted materials without clear compensation or permission challenges traditional notions of creative ownership and control.

Those interested in how AI is evolving and impacting creative professions may find it useful to explore up-to-date AI courses that cover these developments and offer practical insights.