Perplexity AI revenue share model appeases publishers
Publishers earn a share of the startup's revenue when their work is referenced. They also have access to the vendor's LLM API and developer support. But writers are not included.
Following allegations that its generative AI system was infringing on news articles and illegally scraping content from the web, Perplexity AI has introduced a revenue sharing model.
The AI startup, whose AI search engine that provides links to sources has quickly become popular, introduced its Publishers' Program on July 30.
The AI startup started the program by partnering with Time, Der Spiegel, Fortune, Entrepreneur, The Texas Tribune and WordPress.
Perplexity said it will introduce advertising through its related question feature, which allows brands to pay to ask follow-up questions.
Publishers can then earn a revenue from interactions when their content is referenced.
Partners will also have access to Perplexity's Online LLM API and developer support. This will allow each publisher to create a custom answer engine on its website.
Employees of the publishers working with Perplexity will be able to use the AI vendor's Enterprise Pro offering free for a year.
The new revenue-sharing model comes after Forbes said the vendor's generative AI system purposely infringed on the publication's paywalled original reporting.
Avoiding a lawsuit
Perplexity's latest move appears, in some ways, to be an approach that will let it avoid the situation its competitor, OpenAI, is facing in its copyright infringement battle with The New York Times.
"They're trying not to get sued," said Juliette Powell, an author and New York University adjunct professor who writes and teaches about AI.
Since the legal threat the vendor faces has been growing dramatically, Perplexity needed to show that it takes threats of legal action seriously, said Michael Bennett, AI policy adviser at Northeastern University.
"They were modifying their actions to neutralize what content creators see as a threat to their viability, their commercial viability," Bennett said.
Perplexity also wants to distinguish itself from OpenAI, he added.
They're trying not to get sued.
Juliette PowellProfessor, NYU
OpenAI offers a revenue cut to some publishers, notably the Associated Press.
However, compared to OpenAI, Perplexity's publishing program includes a broader range of content creators and publishers, Bennett said.
For example, not only will publishers be paid for the content that contributed to the initial answer for the user, but the user can also go to the publishers' website, which is a value.
This route might also be a way for Perplexity to remain relevant and grow in a competitive market in which the company might soon get bought out, said Steven Dickens, an analyst with Futurum Group.
"There's a lot of other well-funded players that they've got to compete with," Dickens said. "I don't think their money is going to come from the consumer piece. It's going to come from the APIs and enterprise piece."
Shortchanging writers
However, using the publishers' program to grow only benefits Perplexity and the publishers, not individual content creators or writers, Powell said.
"The rev share goes to the media companies. But those media companies aren't necessarily going to be sharing that revenue with their writers," she said. "They have absolutely no reason to because a lot of these writers -- a lot of the actual creators of the content -- are under contract but don't even own their own stuff."
It would be different matter if Perplexity or the publishers included a clause that the content creators would get a share of the revenue from their work, she added.
However, Perplexity does not have a responsibility to do that because its focus is on getting a bigger market share, Powell said.
"No matter how you slice it, the actual human creators are the ones that are getting the short stick," she said. "That's the piece of the story that the creators themselves are aware of but don't know what to do about."
Esther Ajao is a TechTarget Editorial news writer and podcast host covering artificial intelligence software and systems.
