Bill Gross’s ProRata Aims to Solve Generative AI’s Copyright Crisis with Innovative Pay-Per-Use Model

To review this article again, go to My Profile and then check your saved stories.

Innovative AI Faces a 'Theft' Issue, and This Entrepreneur Has a Solution

In the tech industry, Bill Gross became well-known during the 1990s for devising an inventive method for search engines to profit from advertisements. His strategy involved advertisers paying whenever their ads were clicked on. Today, the individual behind the "pay-per-click" concept has established a new venture named ProRata. This company is introducing a bold, perhaps overly ambitious, business approach: "AI pay-per-use."

Gross, the chief executive of the Pasadena-based firm, speaks bluntly about the generative AI sector. "It amounts to theft," he asserts. "They are essentially pilfering and repurposing global knowledge for their own gain."

AI firms frequently claim that to develop advanced generative technologies, they require large amounts of data, asserting that harvesting data from the internet—be it text from websites, videos or captions from YouTube, or books sourced from unauthorized libraries—is permissible by law. Gross, however, doesn't accept this rationale, dismissing it as nonsense.

Numerous executives in the media industry, alongside creators such as authors, artists, composers, and other copyright owners, are taking a stand. They find it challenging to stay abreast of the continuous stream of legal actions being taken against artificial intelligence firms, accusing these companies of essentially engaging in piracy through their operational methods.

Gross believes that ProRata presents a superior alternative to courtroom disputes. "My goal is to ensure fairness," he states. "In my view, resorting to legal action isn't the right way to resolve this."

Gross's enterprise is focused on establishing profit-sharing agreements, ensuring that both publishers and individuals receive compensation when their creations are utilized by AI firms. Gross describes the process as follows: "Our approach involves deconstructing the products of generative AI, be it written content, visuals, tunes, or films, into their elemental parts. This allows us to trace their origins, allocate a proportional share of credit to each copyright owner, and compensate them as a result." ProRata has applied for patents for its innovative algorithms designed to accurately attribute sources and manage compensations.

This week marked the debut of a company that successfully secured $25 million in funding, introducing collaborations with prominent entities such as Universal Music Group, the Financial Times, The Atlantic, and Axel Springer, a major media conglomerate. Furthermore, it has established agreements with well-known authors like Tony Robbins, Neal Postman, and Scott Galloway, who have significant audiences. (The company has also engaged in a partnership with Anthony Scaramucci, who previously served as the White House communications director.)

Journalism expert Jeff Jarvis, despite his view that using web data for AI development falls under fair use, has joined the cause. Speaking to WIRED, he mentioned that it's a wise move for those in the media sector to unite in order to provide AI firms with access to "reliable and up-to-date information" for their content generation. "I'm optimistic that ProRata could initiate conversations that may lead to the creation of APIs [application programming interfaces] for different types of content," he states.

After the company made its first announcement, Gross received an overwhelming number of inquiries from various companies eager to join, among them a message from Time’s CEO, Jessica Sibley. Time magazine has indeed finalized an agreement with ProRata, as confirmed to WIRED by the publisher. Gross is aiming to secure contracts with notable YouTubers and other prominent online personalities.

The focal point to understand is "plans." The business is just getting off the ground, and Gross is making ambitious claims. As an initial demonstration, ProRata is rolling out its subscription-based chatbot-like search engine in October. Differing from other AI search offerings, ProRata's search feature will solely rely on officially licensed data, avoiding anything gathered through web scraping. "Nothing from Reddit," he emphasizes.

Ed Newton-Rex, previously associated with Stability AI and currently at the helm of the ethical data licensing organization Fairly Trained, is encouraged by the launch of ProRata. "It's uplifting to witness a generative AI firm taking the initiative to license training data prior to launching their model, diverging from the practices of numerous other firms," he remarks. "The agreements they've secured also highlight the willingness of media companies to collaborate with reputable entities."

Gross asserts that a focus on the quality rather than the quantity of data will enhance the search engine's performance, arguing that restricting the model to reliable sources of information will reduce inaccuracies. He maintains, "Having 70 million high-quality documents is, in fact, better than possessing 70 billion of poor quality. This approach will result in more accurate responses."

Furthermore, Gross believes he can attract a sufficient number of subscribers to this AI search engine, which exclusively uses licensed data, to generate the necessary revenue to compensate its data contributors with their designated portion. "Each month, the partners will receive a report from us detailing the search queries, the utilization of their content, and their proportionate share of the earnings," he states.

Several emerging companies are vying for a leading position in the burgeoning industry of licensing training data, notable among them are platforms such as TollBit and Human Native AI. To advocate for enhanced standardization in licensing practices, a nonprofit known as the Dataset Providers Alliance was established earlier this year, with founding participants encompassing entities like the Global Copyright Exchange and Datarade.

The core strategy of ProRata involves offering its proprietary attribution and payment systems for use by other firms, particularly those prominent in the AI sector. Several of these businesses are already forming partnerships directly with content creators. (For example, The Atlantic and Axel Springer have established arrangements with OpenAI.) Gross is optimistic that AI enterprises will see the financial benefit in adopting ProRata's technology over developing similar solutions themselves.

Gross states, "I'm open to providing a license for the system to any interested party. My goal is to set the price so low it's comparable to the fees of Visa or Mastercard."

Recommended for You…

Direct to your email: A selection of our top stories, curated daily especially for you

A flawed update from CrowdStrike brought global computer systems to a halt

The Major Headline: When could the Atlantic Ocean potentially split apart?

Introducing the age of excessive online consumption

Olympic Games: Stay updated with our comprehensive coverage from Paris this season right here.

Additional Content from WIRED

Critiques and Manuals

© 2024 Condé Nast. All rights reserved. Purchases made through our website could result in WIRED receiving a share of the sales, thanks to our affiliate agreements with various retailers. Content from this site cannot be copied, shared, broadcast, stored, or used in any form without explicit written consent from Condé Nast. Advertising Choices

Choose a global website