Over the past few years, AI technology has progressed at a rapid pace.

This includes large language models, which are typically trained on a broad datasets of text; the more, the better.

When AI hit the mainstream, it became apparent that rightsholders were not always pleased to learn that their works had been used to train AI. Those works include millions of books and articles, as well as libraries of music and photographs.

There are currently dozens of lawsuits in U.S. courts with this dispute at their core. The AI companies typically take the position that using copyrighted works as model input is fair use, but rightsholders disagree. They want to be compensated for use of their works.

With many billions of dollars at stake, this legal question may ultimately end up at the Supreme Court. Alternatively, lawmakers could address the topic in targeted bills , which could support either side of the debate.

Trump: Paying for All Copyrighted Content Is Not Doable

Yesterday, President Trump weighed in on the issue during the “Winning the AI Race” summit, where he suggested that it is “not doable” for AI companies to pay for all copyrighted content used to train AI models.

“You can’t be expected to have a successful AI program when every single article, book or anything else that you’ve read or studied, you’re supposed to pay for. You just can’t do it because it’s not doable,” Trump said.

“When a person reads a book or an article, you’ve gained great knowledge. That does not mean that you’re violating copyright laws or have to make deals with every content provider,” the president added.

AI-related copyright issues are complex and involve both content used for input and content output by models. In his comments, President Trump appears to refer mostly to the former. While some AI companies have made licensing deals, many have used unauthorized libraries often derived from pirate sites.

China and Anna’s Archive

One of these so-called shadow libraries, Anna’s Archive, previously said that high-speed access to data had already been provided to dozens of AI companies and data brokers, most of which are based in China.

Anna’s Archive warned that if countries want to stay relevant in the AI race, copyright should not be a restricting factor. “It’s a matter of national security,” Anna’s Archive said.

President Trump is also aware of various stances that other nations have taken on copyright and AI. In his speech, he explicitly suggested that paying for all copyrighted material is not a viable option, noting that “China’s not doing it.”

“You have to be able to play by the same set of rules,” Trump said, adding that it doesn’t mean that AI should be able to reproduce copyrighted works.

“Of course, you can’t copy or plagiarize an article, but if you read an article and learn from it, we have to allow AI to use that pool of knowledge without going through the complexity of contract negotiations, of which there would be thousands for every time we use AI.”

Acknowledging Opposition

Trump said a number of people warned him that this position would not be popular with everyone, but it is necessary to advance AI.

“I was told before I got up here, this is an unpopular thing because some people, they don’t want that. But I want you to be successful,” he said.

These comments come after Meta and Anthropic celebrated early ‘fair use’ victories in the U.S., but there’s still a long way to go before this complex topic is completely settled before the courts. Whether the U.S. administration plans to take steps to advance concrete legislation on the issue is not immediately clear.

A final fair use ruling won’t necessarily resolve all AI copyright battles either. Meta and other companies have also been accused of actively distributing pirated books as part of their information-gathering process, which is a separate copyright infringement issue.

From: TF , for the latest news on copyright battles, piracy and more.