As AI companies swarm to leverage other media outlets' content for training their advanced language models, an emerging sect of intermediaries is emerging to negotiate licensing deals between content creators and these tech firms.
These Entrepreneurial Ventures Ensure that AI Businesses Fork Over Compensation for Utilizing Content
Traditionally, publishers coping with AI startups mining their content to bolster powerful large language models have two alternatives: they can challenge the company for copyright infringement lawsuits or formulate a single-time licensing agreement for their archives.
However, a fresh wave of businesses is offering a third alternative, declaring to help publishers recoup some earnings when their work is utilized or cited by AI, reimbursing them for lost page views.
One such company, TollBit, performs as a digital toll booth of sorts, imposing a fee on AI companies whenever they extract a publisher’s content. Another organization, ProRata, develops technology to aid AI companies in compensating publishers based on the degree of their content present in AI-generated outputs. Then, there's ScalePost, which is building a repository of licensed content that AI businesses can access by paying a fee.
Currently, AI titans like OpenAI, Anthropic, and Perplexity, which repurposed an exclusive, paywalled investigation from "Our Website" across various platforms earlier this year, are renowned for disregarding their own protocols to curb web crawlers from scraping content. This defiance has resulted in high-profile lawsuits, with entities like the New York Times and Dow Jones asserting that unauthorized scraping violates copyright law. Other media outlets have opted to collaborate instead of engaging in litigation. For instance, OpenAI is paying DotDash Meredith at least $16 million annually to license its content based on Adweek's reporting. Moreover, Thomson Reuters earned $33 million in year-to-date revenue from AI content licensing deals in its latest quarterly earnings report.
As AI systems digest content in real-time to provide up-to-date information, and AI-search engines become increasingly popular, publishers are troubled by the outflow of revenue-generating traffic from their sites.
"There's no secret that publishers are struggling right now, so we were hunting for opportunities to reap the value for our content that it deserves," Burhan Hamid, CTO of Time, explained to Our Website. Time is one of roughly 400 firms, including Adweek and Hearst Corporation, that have registered to use TollBit, with Hamid attributing additional insights into the bots attempting to access Time’s content to this service.
AI leaders in this domain, such as OpenAI CEO Sam Altman and Google CEO Sundar Pichai, have acknowledged the necessity of devising a new monetary model for rewarding creators in the context of AI scraping. During the New York Times’ Dealbook Summit in December, Altman asserted that micropayments might function as a potential method. "We need to find new revenue streams for creators," Altman explained. OpenAI remained unresponsive to our request for comment.
"There will be a marketplace in the future, I think. There will be creators who create content exclusively for AI models and earn a living from it," Pichai said at the Dealbook Summit. Google also declined to comment on the matter.
"We can calculate, ‘Your article was referenced in this AI response about 33%, so you deserve 33% of the shared 50% revenue.’"
AI companies can utilize TollBit's platform to gain access to publishers' archives, filtering out unauthorized content or material not licensed for use. TollBit provides media companies with data analytics on how frequently bots are scraping their sites and sets up a 'bot paywall' that redirects web scrapers to a different page warning them that they do not have permission to access the content. TollBit charges AI companies a transaction fee and offers them a marketplace of licensed data and a control panel to manage it. CEO Toshit Panigrahi, who cofounded the company with Olivia Joslin in 2023, confirms that TollBit is currently in negotiations with major AI companies but does not disclose their names.
"We need to simplify this machine-to-machine interaction without involving a human," Panigrahi conveyed to Our Website.
Investor interest in this nascent space is still modest in comparison to the billions of dollars that have flowed into AI startups themselves and companies offering them compute or processing power. TollBit has managed to secure approximately $30 million in venture capital financing from Lightspeed Venture Partners and others. ProRata sealed a $25 million Series A round last month.
Publishers recognize the compelling need for such technology. Mike Beyman, head of strategy and operations at Adweek, hopes that TollBit will facilitate the availability of his firm’s content for AI builders who may not possess the resources to negotiate individual licensing agreements with multiple publishers. “My mission is to eradicate barriers to entry for potential partners and ensure fair, mutually acceptable compensation for the use of our content,” Beyman affirmed.
However, allocating a price tag to the content utilized in training AI models is a formidable and intricate task. Patrick Hainault, a VP at Mansueto Ventures, the publisher of magazines Fast Company and Inc, insists that factors like an article’s originality and relevance should also influence the pricing. “Pricing is an alchemy under the best of circumstances, and in this revolutionary landscape, there will be a great deal of experimentation,” Hainault clarified to Our Website.
Bill Gross, the headhoncho of ProRata, claims to have cracked the code for AI search engines. He's come up with a system he calls an "attribution percentage" - a method for determining how much of an AI-generated response stems from a specific content source and how much revenue should be doled out to them. ProRata's tech wizards design an algorithm that splits up the AI's output into various segments, pinpointing and rating the most distinctive insights found in AI-generated responses, thereby compensating the content creators appropriately.
Gross explained, "We can figure out, 'Hey, your article contributed to this AI response for around 33% of the answer, so you get 33% of the 50-50 revenue share.'"
Recently, ProRata introduced their own AI-powered search engine named Gist.ai, which partners with around 400 media heavyweights like Fortune, The Atlantic, and The Financial Times to utilize only licensed content. At the end of each month, publishers receive a breakdown of their content's usage in AI-generated responses along with compensations for their contributions.
Competing companies such as ScalePost AI are also making strides in providing AI access to publishers' video and audio content. By brokering monetization deals between AI firms and media organizations, ScalePost offers a platform for US and international media companies to monitor bot traffic for each URL. "We are the reliable source of information regarding the happenings at the link level," ScalePost CEO Ahmed Mallik shared with Our Website. (Disclosure: Tollbit, ProRata, and Scale Post are under discussion with Our Website.)
ScalePost also acts as a bridge between publishers and AI founders, pairing them through community events. In July, they joined forces with AI search startup Perplexity on its revenue-sharing program, allowing publishers to sponsor relevant questions and collect a share of Perplexity's earnings on answers that cite the publisher as a source. ScalePost shares data with participating publishers, revealing how their content is featured within Perplexity's AI-generated responses. Perplexity praise ScalePost as a "valuable partner," assisting in introductions and expanding its revenue-sharing program to around 20 publishers.
Furthermore, ScalePost introduced a "bot module" that tags and categorizes approximately 800 AI bots. "Each web visit is accounted for, offering a transparency that publishers are naturally interested in," Mallik told Our Website. "If we can track them, we can block them."
Sarah Emerson contributed reporting.
- TollBit, a company offering a third alternative for publishers, operates as a digital toll booth, charging AI companies a fee whenever they extract content from licensed publishers.
- ProRata, another organization, develops technology to help AI companies compensate publishers based on the degree of their content present in AI-generated outputs.
- ScalePost, which aims to build a repository of licensed content, allows AI businesses to access this content by paying a fee, helping publishers recoup some earnings.
- In the realm of content licensing, OpenAI, Anthropic, and Perplexity, among other AI titans, have been repurposing content from various media outlets without proper authorization, leading to high-profile lawsuits.
- Apart from TollBit, ProRata, and ScalePost, there are other companies like ScalePost AI that are brokering monetization deals between AI firms and media organizations, providing publishers with a way to monitor and block unauthorized bot traffic.
