The NYT's cease and desist move against Perplexity AI
The AI startup is facing accusations it took content from a publisher without permission. This is the second time the NYT has gone after a vendor for misappropriating its content.
Perplexity AI has two weeks to respond to a cease-and-desist letter from newspaper publisher The New York Times.
In the letter, the Times accuses the generative AI search vendor of accessing and using its content without authorization to create content such as summaries and other outputs that infringe on copyright laws, according to multiple media outlets.
The New York Times is not the only publisher that has called out the AI startup for scraping content online.
A track record
In June, Forbes threatened legal action against Perplexity, accusing the startup of willful infringement by stealing text and images.
Through the program, Perplexity partners with publishers such as Time, Fortune, Entrepreneur, The Texas Tribune and WordPress.com. Publishers can earn revenue when their work is referenced.
Perplexity is also not the first AI vendor to get in trouble with the NYT.
The publisher is in a legal battle with ChatGPT maker OpenAI and its partner Microsoft. The Times accuses OpenAI and the tech giant of infringing on its content.
A different approach
The suit against OpenAI has yet to conclude, so The Times is taking a different approach this time, said Sarah Kreps, John L. Wetherill professor of government and law at Cornell University.
"Cease and desist approaches are less confrontational, less expensive and faster," Kreps said. "[There is] more of an opportunity for negotiation than a lawsuit, which seems preferable given the circumstances of novel technology with unwritten interpretations of copyright laws."
The cease-and-desist letter is also a smart move by The Times, Northeastern University director of education curriculum and business lead for responsible AI Michael Bennett said.
While the current lawsuit against OpenAI is in open litigation, if The New York Times wins, it could force AI vendors like Perplexity to negotiate a contractually bound financial arrangement, Bennett said.
If The Times doesn't prevail, it loses much of the leverage it had before the court made a decision, Bennett added.
"If things go their way, then great, they're golden," he said. "If it does not go their way, then they will have lost that leverage. It probably behooves them now to move quickly, send out these cease-and-desists when they've got a reasonable basis for doing so."
Moreover, the media giant's move sends a warning to other AI vendors that The Times is serious about protecting its intellectual property, said Vincent Allen, a partner at the intellectual property law firm Carstens, Allen & Gourley, LLP.
"Ultimately what the Times is after is [making] money on their content," Allen said. "Right now, the Times feels like … these works are being used, and it's diverting traffic away from their paid website."
Facts vs. creative expression
On Perplexity's side, the AI startup has responded to The Times' claims by arguing that facts can't be copyrighted.
The vendor said in a statement, "We aren't scraping data for building foundation models, but rather indexing web pages and surfacing factual content as citations to inform responses when a user asks a question. The law recognizes that no one organization owns the copyright over facts. This is what allows us to have a rich and open information ecosystem, not to mention, it gives news organizations the ability to report on topics that were previously covered by another news outlet."
Perplexity's argument is credible, and the AI startup is more than summarizing the facts, Kreps said.
Cease and desist approaches are less confrontational, less expensive and faster.
Sarah KrepsProfessor, Cornell University
"It's indexing web pages to provide factual answers, which is no different from what an AI-powered Google search does," she said. For example, when a user puts in a query about the NYT lawsuit against OpenAI in Google search, the engine provides summaries through its AI overviews and links.
"If Perplexity is culpable then the entire AI industry will be culpable as well," Kreps added. "I do not see much difference between Perplexity's model and Google's AI-assisted search engine."
She added that the cease-and-desist could become a barometer of fair use for the AI industry. Fair use is the doctrine that use of some copyrighted work is allowed under freedom of expression.
"Perplexity's approach is far more representative of how other AI platforms are operating than OpenAI's ChatGPT, which was far more intrusive and ambiguous in terms of compliance with copyright," she continued.
She added that Perplexity is more like Google in explaining facts and giving references, while ChatGPT doesn't say where it gets its information.
"It doesn't give credit to sources and is obviating the need for additional searches because it's comprehensive but also because it doesn't give the user links for further research," Kreps continued.
However, the NYT has gone to great lengths with the cease-and-desist to show that the content Perplexity allegedly took has to do with the publisher's creative expression of facts, Bennett said.
"The New York Times is subtly saying in its cease-and-desist, 'We don't publish facts. What we publish is an expressive interpretation of the facts,'" he said. "And so copyright does protect them."
The Times is also seeking more information about how Perplexity is getting its data.
If Perplexity is scraping data from The Times' website, that probably means that the vendor is making copies of individual articles to put it into their system and then summarize them, Allen said.
"That might be where the copyright infringement claim would come in," he said. "Nobody's really for certain how it's all going to play out. But it is a big deal for the content providers because they want to make sure that they get compensated for the work that they created."
Esther Ajao is a TechTarget Editorial news writer and podcast host covering artificial intelligence software and systems.
