Reddit has initiated legal action against Anthropic for purportedly utilizing the platform’s data to develop artificial intelligence models without obtaining a proper licensing agreement. The lawsuit was filed in a Northern California court and asserts that Anthropic’s unauthorized use of Reddit content for commercial gain violates both legal standards and Reddit’s user agreement. This lawsuit marks Reddit as the first major technology company to challenge an AI model provider regarding the practices surrounding training data, aligning itself with various publishers who have taken similar actions against tech companies.
For instance, The New York Times has sued OpenAI and Microsoft for leveraging its news articles without proper compensation or authorization. Additionally, notable authors like Sarah Silverman have filed lawsuits against Meta for using their written works to train AI systems without consent. Music publishers and artists have also raised concerns over the unauthorized use of their content by AI-focused startups.
Ben Lee, Reddit’s chief legal officer, emphasized that the company will not permit entities like Anthropic to profit from Reddit content while disregarding the rights and privacy of its users. Notably, Reddit has previously established agreements with other AI companies, including OpenAI and Google, to allow them to utilize its data for training AI models under specific terms that safeguard user interests. In the legal filing, Reddit alleges that it had clearly communicated to Anthropic that they lacked authorization to scrape or utilize Reddit content, but Anthropic reportedly declined to engage in discussions.
Anthropic has responded by stating that it disagrees with Reddit’s allegations and intends to defend itself robustly. Reddit also claims that Anthropic’s scraping activities disregarded established protocols meant to prevent automated systems from crawling specific content on its site. The platform is seeking compensatory damages, restitution for the gains Anthropic has made through scraping its content, and an injunction to stop further use of its material.