Copyright Rule Would Ease News Registration as AI Fight Looms

A proposal to ease copyright registration for online news publishers could encourage more outlets to register their work and open the door for more lawsuits as the journalism industry grapples with artificial intelligence companies training models on their content.

The proposed rulemaking from the US Copyright Office would allow publishers with frequently updated news websites to register up to a month’s worth of updates at a time using representative examples, rather than registering individual updates or stories. That would eliminate a “virtual bar” to registration publishers have criticized for more than a decade, which has led many digital outlets to shun seeking copyrights.

“The current state of affairs really handicaps digital media companies,” said Edward Klaris, managing partner of Klaris Law PLLC.

Copyright protection starts as soon as a work is created, but registration gives the owner the right to sue for statutory damages. The New York Times, which regularly registers its work, is seeking statutory damages in its lawsuit against OpenAI and Microsoft and cites its “millions of articles that have registered copyrights.”

Publishers have a “difficult time protecting their sites,” Klaris said, and many have simply opted against navigating the onerous copyright system. That leaves them vulnerable to infringement, including from AI companies scraping their content to train their large-language models. And if they register their news stories, photos and videos later in anticipation of filing a lawsuit, they are precluded from seeking statutory damages, he said.

The proposed rule would “make it much easier for publishers of digital content, and much more appealing for them to protect their work,” he said.

Danielle Coffey, President and CEO of the press trade association News/Media Alliance, stressed the importance of copyright protection for the journalism industry in light of AI’s proliferation at a Senate Judiciary subcommittee hearing Wednesday. Large tech companies, which are the dominant distributors of news content, are responsible for the news industry’s economic decline, she said.

“This marketplace imbalance will only be increased” by AI, Coffey said.

Rulemaking Details

The Copyright Office’s current process requires publishers with dynamic news sites to submit a separate application, deposit, and filing fee for each website update they’re seeking to register. For websites that update constantly in a fast-paced news environment, that means possibly dozens or more updates each day, all of which would require separate registrations.

Allowing those publishers to submit a month’s worth of updates is significant just from a cost perspective, said Thomas Magnani, an Arnold & Porter partner, because “once a day is still one filing fee per day.”

The proposed changes would also allow publishers to register more content in one application using representative examples, which means they wouldn’t have to be as selective about what specific works they register, Magnani said.

That change could “embolden online news publishers to enforce those copyrights,” he said.

Gannett Co. Inc., which publishes USA Today and hundreds of local news pages throughout the country, praised the proposed changes in a statement.

“If adopted, the proposal would streamline the copyright registration process for online newspapers,” a Gannett spokesperson said. “This would be significant for Gannett, as the largest publisher in the nation, as well as for all other online news publishers, who presently face significant burdens with registering the copyrights in their original works of authorship such as text, graphics and photographs.”

Santa Clara University School of Law Professor Tyler Ochoa said in an email he expects the rule to be adopted with few, if any, changes.

“It will be of some assistance to news publishers in bringing copyright lawsuits against alleged infringers, including news aggregators and generative AI companies,” Ochoa said.

“Anything that makes registration easier will have a marginal effect of making it easier for copyright owners to bring a lawsuit and to obtain those remedies,” he said. But, he added, “I seriously doubt that we will see a rash of new lawsuits as a result.”

Klaris, on the other hand, said he would expect more lawsuits. Copyright holders can only seek statutory damages—which are calculated for inch individual infringement—if they register the work prior to the infringement or within three months of its publication.

“People have copyright interest in their work without registering,” he said. “What the registration does is it gives people the right to sue for statutory damages, which is oftentimes higher than the actual damages. And that’s just an easier number to assert.”

The threat of statutory damages could also be a “big lever” to prevent copying, said Knobbe Martens partner Mauricio Uribe. Statutory damages are based on the number of works infringed, so the more works infringed the more damages the copyright holder can pursue.

“The ease of the administrative burden is definitely much less, or certainly outweighed by the benefit you get,” he said of the new rulemaking.

The Copyright Office clarified that to qualify for the new registration process, a site must “be a primary source of written information on current events, either local, national, or international in scope, that contains a broad range of news on all subjects and activities and is not limited to any specific subject matter.” That scope raises the question of whether trade publications would qualify, both Klaris and Magnani said.

“It does seem kind of absurd to knock out people like news publishers that have a specific type of content, whether it’s real estate news, tax news,” said Klaris. “It should be more broadly applicable.”

Uribe also said limiting the registration option to websites that have URLs under a particular domain name could age poorly given how quickly technology evolves.

What about “X or any other kind of social media that ultimately could end up being five years from now the way we consume it in a very different way?” he said. “Or some other medium, or some other device, and the traditional URL goes by the wayside?”

Comments on the rulemaking are due to the Copyright Office on Feb. 20.

Licensing Negotiations

Even if a publisher’s work isn’t registered, it can still be licensed. OpenAI struck a multi-million dollar deal with Axel Springer SE, which owns publications including Politico and Business Insider, in December to use its articles to develop AI models. Axel Springer, Politico, and Business Insider appear to collectively have less than a dozen registered news articles in the Copyright Office’s public records system. OpenAI is also reportedly in negotiations with CNN, Fox Corp., and Time to license their news content, people familiar told Bloomberg News.

Even when news sources aren’t copyrighted, Uribe said, there “still would be value in terms of getting rights to that.”

A licensing agreement can represent “peace and zero risk,” Uribe said, and have political benefits. He noted that Adobe Inc., whose AI tool Firefly has been referred to as “copyright-safe” because it’s trained on its own stock images, openly licensed content, and other public domain content, seemed to be treated differently than other companies at a Senate hearing on AI this past summer. “The fact is that there is optics and politics and business kind of that that gets factored into there.”

Media executives on Wednesday urged Congress to craft legislation clarifying that the use of unlicensed copyrighted content to train AI models shouldn’t fall under the ‘fair use’ doctrine, which allows unapproved copying in certain scenarios. Roger Lynch, Condé Nast CEO, said AI companies are required by law to license publisher content for their generative AI tools.

“Currently deployed Gen AI tools have been built with stolen goods,” Lynch said.

Clarification from Congress would enable private negotiations, he said, and the free market will take care of the rest. “Although they negotiate with us, their starting point is ‘We don’t want to pay for content that we know that we should be able to get for free.’”

Though publishers with registered content can pursue litigation against AI companies if licensing talks don’t pan out—as happened ahead of the New York Times’ suit—those fights will still face major challenges.

“If it’s an arms race of who can spend the most on litigation, we know that the tech industry beats out everyone else,” National Association of Broadcasters President and CEO Curtis LeGeyt said.