OpenAI and the White House have accused DeepSeek of using ChatGPT to cheaply train its brand-new chatbot.

- Experts in tech law state OpenAI has little recourse under copyright and contract law.

- OpenAI's terms of use may use but are mostly unenforceable, they state.


This week, OpenAI and pl.velo.wiki the White House implicated DeepSeek of something comparable to theft.

In a flurry of press declarations, they stated the Chinese upstart had actually bombarded OpenAI's chatbots with queries and hoovered up the resulting data trove to rapidly and cheaply train a design that's now nearly as great.

The Trump administration's leading AI czar said this training process, called "distilling," amounted to copyright theft. OpenAI, on the other hand, informed Business Insider and other outlets that it's examining whether "DeepSeek may have inappropriately distilled our designs."

OpenAI is not saying whether the company plans to pursue legal action, instead promising what a spokesperson called "aggressive, proactive countermeasures to protect our innovation."

But could it? Could it sue DeepSeek on "you took our content" premises, much like the grounds OpenAI was itself sued on in a continuous copyright claim filed in 2023 by The New York City Times and other news outlets?

BI presented this question to professionals in innovation law, who stated tough DeepSeek in the courts would be an uphill struggle for OpenAI now that the content-appropriation shoe is on the other foot.

OpenAI would have a difficult time showing an intellectual residential or commercial property or copyright claim, these legal representatives said.

"The concern is whether ChatGPT outputs" - implying the responses it generates in action to questions - "are copyrightable at all," Mason Kortz of Harvard Law School said.

That's due to the fact that it's unclear whether the answers ChatGPT spits out certify as "imagination," he stated.

"There's a doctrine that states creative expression is copyrightable, but realities and concepts are not," Kortz, oke.zone who teaches at Harvard's Cyberlaw Clinic, said.

"There's a huge question in intellectual property law today about whether the outputs of a generative AI can ever make up imaginative expression or if they are always unguarded realities," he included.

Could OpenAI roll those dice anyway and declare that its outputs are secured?

That's not likely, the legal representatives said.

OpenAI is currently on the record in The New York Times' copyright case arguing that training AI is an allowed "reasonable usage" exception to copyright protection.

If they do a 180 and tell DeepSeek that training is not a fair usage, "that might come back to type of bite them," Kortz said. "DeepSeek could say, 'Hey, weren't you simply saying that training is reasonable usage?'"

There may be a difference in between the Times and DeepSeek cases, Kortz added.

"Maybe it's more transformative to turn news posts into a model" - as the Times implicates OpenAI of doing - "than it is to turn outputs of a model into another model," as DeepSeek is said to have actually done, Kortz stated.

"But this still puts OpenAI in a quite predicament with regard to the line it's been toeing relating to reasonable use," he added.

A breach-of-contract claim is more most likely

A breach-of-contract suit is much likelier than an IP-based claim, though it includes its own set of issues, said Anupam Chander, who teaches technology law at Georgetown University.

Related stories

The terms of service for Big Tech chatbots like those established by OpenAI and Anthropic forbid using their content as training fodder for a contending AI design.

"So perhaps that's the lawsuit you may potentially bring - a contract-based claim, not an IP-based claim," Chander said.

"Not, 'You copied something from me,' but that you benefited from my design to do something that you were not allowed to do under our agreement."

There might be a drawback, Chander and Kortz stated. OpenAI's terms of service need that the majority of claims be solved through arbitration, not lawsuits. There's an exception for hikvisiondb.webcam claims "to stop unauthorized use or abuse of the Services or copyright violation or misappropriation."

There's a larger drawback, though, experts stated.

"You must understand that the dazzling scholar Mark Lemley and a coauthor argue that AI terms of usage are most likely unenforceable," Chander stated. He was referring to a January 10 paper, "The Mirage of Artificial Intelligence Terms of Use Restrictions," by Stanford Law's Mark A. Lemley and Peter Henderson of Princeton University's Center for Infotech Policy.

To date, "no design creator has really tried to enforce these terms with financial penalties or injunctive relief," the paper states.

"This is most likely for great reason: we believe that the legal enforceability of these licenses is questionable," it adds. That remains in part due to the fact that design outputs "are mainly not copyrightable" and due to the fact that laws like the Digital Millennium Copyright Act and the Computer Fraud and Abuse Act "offer minimal option," it says.

"I think they are most likely unenforceable," Lemley informed BI of OpenAI's regards to service, "due to the fact that DeepSeek didn't take anything copyrighted by OpenAI and because courts generally won't impose agreements not to compete in the absence of an IP right that would avoid that competition."

Lawsuits between celebrations in various countries, each with its own legal and enforcement systems, are constantly challenging, Kortz stated.

Even if OpenAI cleared all the above obstacles and won a judgment from an US court or arbitrator, "in order to get DeepSeek to turn over cash or stop doing what it's doing, the enforcement would come down to the Chinese legal system," he stated.

Here, OpenAI would be at the grace of another extremely complicated location of law - the enforcement of foreign judgments and the balancing of individual and corporate rights and nationwide sovereignty - that extends back to before the founding of the US.

"So this is, a long, made complex, fraught process," Kortz added.

Could OpenAI have safeguarded itself better from a distilling attack?

"They might have used technical steps to block repeated access to their site," Lemley said. "But doing so would also hinder typical clients."

He included: "I don't believe they could, or should, have a valid legal claim versus the searching of uncopyrightable information from a public site."

Representatives for DeepSeek did not right away respond to an ask for comment.

"We understand that groups in the PRC are actively working to utilize techniques, including what's known as distillation, to try to duplicate advanced U.S. AI designs," Rhianna Donaldson, an OpenAI spokesperson, junkerhq.net told BI in an emailed statement.