On October 22nd, local time, the US social media platform Reddit filed a lawsuit in a New York federal court, accusing the artificial intelligence startup Perplexity and three data-scraping companies of unauthorized access to its forum data for use in AI model training. Reddit's Chief Legal Officer, Ben Lee, stated in a statement that AI companies are locked in an "arms race" for high-quality content, a pressure that has spawned an industrial-scale "data laundering" economy, with Reddit, due to its vast repository of user conversations, a prime target.
The lawsuit documents indicate that Perplexity obtained Reddit data collected through Google search results from at least one co-defendant without the platform's authorization. This is Reddit's second similar lawsuit this year; in July, the company sued Anthropic for data access violations and previously reached content licensing agreements with Google and OpenAI valued at $203 million. A Perplexity spokesperson responded that the company "stands firmly by the principle of free user access to public knowledge" and that its AI answers are "fact-based and responsible."
Founded in 2022, Perplexity is an AI search unicorn valued at $18 billion, specializing in question-and-answer engine services. Its CEO, Aravind Srinivas, has stated that AI should help users more easily access knowledge. However, Reddit has accused the platform of laundering data sources through third parties, essentially a shady operation to circumvent copyright liability. Legal experts point out that this case could become a key precedent in defining the legality of AI training data, and that Reddit's proposed "dynamic pricing" licensing model could become a new industry standard.
