The 28-year-old founders of six-month-old New York-based startup TollBit think we’re living in the “Napster days” of artificial intelligence. Just as people of a certain generation download digital music, companies rip off large swaths of the internet without paying rights holders. They want TollBit to be the iTunes of the AI world.
“It’s kind of the Wild West right now,” Olivia Joslin, the company’s co-founder and chief operating officer, told Engadget. “We want to make it easier for AI companies to pay for the data they need.” Their idea is simple: create a marketplace that connects AI companies who need access to fresh, high-quality data with publishers who actually spend money.
AI companies have indeed recently started paying for (some of) the data they need from news broadcasters. OpenAI began the arms race in late 2022, but the company signed the first of many licensing agreements with the Associated Press just a year ago. Later that year, OpenAI announced a partnership with German publisher Axel Springer Business Insider and Politics in the United States. Including many publishers Voxthe Financial TimesNews Corp and TIMEit has since signed deals with OpenAI and Google.
But that still leaves countless other publishers and creators out in the cold — without the option to strike this Faustian Bargain, even if they wanted to. This is the “long tail” of publishers that TollBit wants to target.
“Strong AI models already exist and they’re already trained,” Toshit Panigrahi, TollBit’s co-founder and CEO, told Engadget. “There are currently thousands of programs that take these existing models off the shelves. What they need is fresh content. But right now, there’s no infrastructure — neither for them to buy it, nor for content producers to sell it seamlessly.”
Both Joslin and Panigrahi didn’t know much about the media industry. But they both knew how online marketplaces and platforms worked—they were colleagues at Toast, a platform that allows restaurants to manage billing and reservations. Panigrahi watched both deals and lawsuits pile up in the AI sector, then called Joslin.
Their first conversation was about RAG, which stands for Retrieval-Augmented Generation in the world of artificial intelligence. With RAG, AI models first search for information from specific databases (such as scrapable parts of the internet) and use that information to synthesize an answer, rather than simply relying on training data. Services like ChatGPT don’t know current house prices or breaking news. Instead, they usually get that information by browsing websites. The lack of fresh data is why AI chatbots are so common stumbled with queries about breaking news events — if they don’t break the latest information, they simply can’t keep up.
“We thought using content for RAG was a completely different thing than using it for training,” Panigrahi said.
According to some estimates, RAG is the future of search engines. More and more people ask questions on the Internet and expect a complete answer instead of a list of blue links. In just over a year, startups like Perplexity, backed by Jess Bezos and NVIDIA, have burst onto the scene with ambitions to take on Google. Even OpenAI have plans Let ChatGPT be your search engine one day. In response, Google has sprung into action—it now extracts relevant information from search results and presents it as a consistent answer at the top of the results page, a feature it calls AI Insights. (O it doesn’t always work wellbut it seems it’s here to stay).
The rise of RAG-based search engines has publishers quaking in their boots. After all, who would make money if artificial intelligence reads the web for us? After Google rolled out AI Reviews at least one earlier this year report It is estimated that publishers will lose more than $2 billion in advertising revenue because fewer people will have a reason to visit their websites. “AI companies also need continued access to high-quality content and data,” Joslin said, “but unless you find some kind of economic model here, there will be no incentive for anyone to create content, and that will be the end of AI applications, too.
Instead of cutting one-time checks, TollBit’s model aims to compensate publishers on an ongoing basis. Hypothetically, if someone’s content was used in a thousand AI-generated responses, they would get paid a thousand times the price they set and can change quickly.
Each time an AI company receives fresh data from a publisher through TollBit, it can pay a small fee set by the publisher, which Panigrahi and Josley say should be roughly equal to what a traditional pageview would charge a publisher. The platform may also block unregistered AI companies from accessing publishers’ data.
So far, the founders claim to have signed up to a hundred publishers and piloted with three AI companies since TollBit launched in February. They declined to say which publishers or AI companies have signed on so far, citing confidentiality clauses, but did not deny they are talking to OpenAI, Anthropic, Google and Meta. So far, they say, no money has changed hands between AI companies and publishers on their platform.
Until that happens, their model is still a giant hypothetical — even though investors have spent $7 million so far. TollBit’s investors include Sunflower Capital, Lerer Hippeau, Operator Collective, AIX and Liquid 2 Ventures, and more are currently “knocking on its doors,” Joslin claimed. At TollBit in April brought Campbell Brown, general counsel, is a former television host who previously served as Meta’s head of news partnerships for the better part of a decade.
Despite some high-profile lawsuits, AI companies yet breaking the internet for free and big to get away from it. Why are they willing to pay publishers for this information? There are three big reasons, the founders say: Since generative artificial intelligence has become widespread, more websites are taking steps to prevent their content from being taken down, which means breaking the internet it becomes difficult and more expensive; nobody wants to deal with ongoing copyright lawsuits; and most importantly, being able to easily pay for content when needed allows AI companies to reach out to smaller and more niche publications, as it is not possible to enter into individual licensing agreements with each website. Joslin also pointed out that many TollBit investors have also invested in AI companies, where they worry they could face lawsuits for using unauthorized content.
Getting AI companies to pay for content could provide a recurring revenue stream not just for big publishers, but for anyone who publishes anything online. Last month, Perplexity was accused of illegally ripping off content Forbes, Wired and Condé Nast — Launched a Publishers Program, under which it plans to share a portion of revenue with publishers if it uses their content to generate answers with AI. The success of the app depends on how much money Perplexity makes when it serves ads on the app later this year. Like Tollbit, another complete hypothetical.
“Our thesis with TollBit is that if you’re losing pageviews today, you should be compensated immediately, not years after a tech company develops an adware,” Panigrahi said of Perplexity’s initiative.
Despite all the existing licensing agreements and technical advances, AI-powered chatbots still make for terrible news sources. They still make up facts and confidently make whole connections to stories that don’t actually exist. But tech companies are now stuffing AI chatbots into every crevice they can, which means many people will still be hearing from one of these products in the near future.
A more cynical take on TollBit’s premise is that the startup is effectively offering hush money to publishers whose job it is to avoid turning out to be misinformation. Its creators, of course, do not agree with its characterization. “We are focused on the AI partners we have on board,” Panigrahi said. “These companies pay a lot of attention to the quality of the input material and the accuracy of the answers. We see that paying for content, even nominal amounts, creates an incentive to respect the raw materials that go into their systems, rather than treating it as a free, fungible commodity.”
