It’s well known that the AI industry rests on shaky legal ground. Companies like OpenAI have built their multi-billion-dollar businesses on the backs of vast tranches of training data, much of which is sourced from copyrighted content. The creators of that content know they’re being ripped off and, more and more, it’s leading to lawsuits. We got another reminder of this conundrum this week, when Reddit sued Anthropic over its use of Redditors’ posts in its training data.
Reddit’s lawsuit, which was filed Wednesday, accuses the Amazon-backed AI company of breaching its user agreement. “As far back as December 2021…Anthropic was already—without authorization and in direct violation of Reddit’s User Agreement—training Claude on Reddit users” posts, the lawsuit claims.
Anthropic, whose flagship product is the AI chatbot Claude, has tried to position itself as the “good guy” of the AI industry—a company that plays by the rules and advances AI frameworks that are considerate of safety and ethical considerations. But, despite its “white knight” PR, the company has repeatedly run into legal issues that throw its supposedly “ethical” businesses practices into question. This week’s litigation is yet another reminder of that.
The lawsuit accuses Anthropic of unjustly enriching itself while also breaching the platform’s user agreement. The suit claims that the AI company’s bots have visited its website over 100,000 times since 2024.
“This case is about the two faces of Anthropic: the public face that attempts to ingratiate itself into the consumer’s consciousness with claims of righteousness and respect for boundaries and the law, and the private face that ignores any rules that interfere with its attempts to further line its pockets,” the litigation states. It adds that Anthropic “continues to publicly admit that it trains its Al technologies on Reddit content.”
When reached for comment by Gizmodo, an Anthropic spokesperson provided the following statement: “We disagree with Reddit’s claims and will defend ourselves vigorously.”
The war over AI content usage has become one of the industry’s most prominent dilemmas. Platforms and artists are aware that their content is being pilfered for the sake of AI fuel, and they’re firing up the lawsuit machine to fight back. At this point, OpenAI has been sued by so many different people and institutions that it’s hard to keep track of it all—everyone from Sarah Silverman, Ta-Nahisi Coates, George R. R. Martin, and Jonathan Franzen, to the Center for Investigative Reporting, The Intercept, a variety of newspapers (including The Denver Post and the Chicago Tribune), and some YouTubers. The New York Times is currently suing the company on similar grounds.
Reddit has sought to insulate itself from getting ripped off by developing contracts with AI companies that clearly stipulate an exchange of content for money. Last February, Reddit struck a deal with Google that allowed the tech giant to use the content on its platform as AI fodder, so long as the company coughed up $60 million a year. Not long afterward, a similar deal was struck with OpenAI. Anthropic doesn’t seem to have gotten the memo, but it surely will now. More and more, it seems like this is the new model for the AI industry: To quote one of my favorite TV shows, you’re going to have to pay the troll toll if you don’t want to get pounded by a lawsuit. It’s obviously a situation that favors large companies. AI companies with the resources will be able to buy access to large amounts of data to fuel their AI habits. Smaller, lesser-resourced firms will be shit out of luck.
Reddit Sues Anthropic, Accusing It of Illegal Data Use
