Perplexity AI will start sharing advertising revenue with news publishers when its chatbot responds to user questions and displays their content, a move likely intended to appease critics who have accused the startup of plagiarism and unethical web scraping.
Perplexity's VP of business Dmitry Shevelenko told TechCrunch that the company had actually been considering the program back in January, before publishers began making accusations. The business rationale for the publisher program was self-protection: For Perplexity to continue providing accurate answers to users' questions, it needs journalists to keep producing new facts about the world.
“How do we work with publishers,” Shevelenko said, “we're not out to cannibalize or compete with publishers, but we need to do our part to ensure we have vibrant and diverse business models and revenue streams.”
As generative AI continues to change the way people search, publishers are looking for new ways to monetize.
Perplexity's initial publishing partners include Automattic, Der Spiegel, Entrepreneur, Fortune, The Texas Tribune and TIME. As part of the multi-year agreement, publishers will have access to Perplexity's API and developer support to build their own custom answer engines for their sites, and all employees will have access to Perplexity's Enterprise Pro service, with enhanced data privacy and security features.
Perplexity hasn't started showing ads on its platform yet, but it plans to do so within the next few months. Shevelenko said the startup lines up “top-tier brands in every major consumer and B2B category.” So, for example, if a user asks Perplexity when they should go to Tokyo, they might see ads from travel-related brands. Any revenue Perplexity makes from those ads is then shared with the publishers whose content is being used to answer that question.
Perplexity declined to disclose details about its ad revenue share, but Shevelenko said it will be in the “double digits” of percentage points.
Other media outlets, including The Atlantic, News Corp, Financial Times, Dotdash Meredith, Axel Springer and Vox Media, have also signed licensing or product agreements with OpenAI in recent months, despite many of these publications' journalists publicly criticizing the company for stealing their content to train AI models.
“I think user behavior will drive this a lot,” Michael Frazier, vice president of data and operations at Entrepreneur Media, told TechCrunch, noting that social media and Google search have taught them that “you have to meet people where they want to be.”
Frazier said the team at Entrepreneur is building tools to let users interact with the publication's content in a more interactive way, and Perplexity's API will help make that happen.
Perplexity is working with ScalePost.ai, a platform that enables collaboration between publishers and AI companies and provides publishers with AI analytics, including how Perplexity cites their content.
“One of the things we want to learn more about is user intent and behavior and how and where they find us,” Fraser continued. “That would be really valuable and give us insight into how we can better support our readers.”
When asked if he was concerned about Entrepreneur's brand being associated with potentially false answers to users' questions (most AI chatbots are prone to hallucinations, and Perplex is no exception), Frazier said it's always a concern, but that he believes Perplex is taking the necessary steps to curb hallucinations.
Perplexity said publishers participating in its program will not receive preferential treatment in search queries, which is a promise OpenAI makes to its media partners. Asked about how often users click on articles Perplexity cites in response to their queries, Shevelenko said click-through rates are not the essence of the program.
“This is the first time that a technology platform is actually giving revenue share to publishers, and we think that's powerful,” he said. “The reason we're doing this, something that Google never did, is because we don't think the primary value we bring to publishers is traffic. We think it's disingenuous to claim otherwise.”
The main value Shevelenko sees is the ad revenue share and access to Perplexity's API and Pro subscription.
Despite their penchant for cozying up to AI companies, not all publishers are willing to go along with it: The New York Times, Raw Story, AlterNet, and The Intercept have all sued OpenAI for using journalists' work to train ChatGPT without properly attributing or citing sources.
According to The Information, Condé Nast, parent company of The New Yorker, Vogue, and Wired, sent a cease and desist letter to Perplexity earlier this month, demanding that the publisher stop using content from its publications in search results. The move appears to come in the wake of a recent Wired report that Perplexity removed content from pages after providing them the page URLs, despite publishers explicitly asking the AI company not to do so.
Forbes magazine sent a similar letter to startup Perplexity, alleging that the company plagiarized one of its news articles in its beta Perplexity Pages feature.
Perplexity isn’t stopping or cancelling, but instead inviting those publishers to join its publisher program.
“We're excited to partner with them,” Shevelenko said.