Reddit has sued Perplexity AI, accusing it of engaging in “industrial-scale” data theft activities.

The social platform Reddit recently filed a lawsuit against Perplexity AI, accusing it of stealing user data on an “industrial scale”. Reddit claims that Perplexity and its partners have collected content from the platform without permission, using it to train artificial intelligence systems. This action violates users’ privacy rights and intellectual property laws. With the rapid development of AI technology, this case has attracted attention from both the crypto community and the tech industry. It also highlights the need for data compliance, protection of user rights, and ethical considerations regarding AI training. This case could become a key issue in future negotiations between digital platforms and AI companies.

In short,

● Social media platform Reddit has sued Perplexity AI, accusing the company of using a “industrial-scale” approach to collect user-generated content.

● Reddit claims that billions of search results were obtained by exploiting tools that bypassed its and Google’s security measures.

● The lawsuit targets Perplexity, SerpApi, Oxylabs, and AWM Proxy as defendants.

Social media platform Reddit filed a lawsuit against Perplexity AI in federal court on Wednesday, accusing the AI company and its data partners of planning a “industrial-scale” operation to collect user-generated content from the platform.

Reddit claims that the other defendants—SerpApi, Oxylabs, and AWM Proxy—have developed and sold tools designed to break down security measures used to protect content, thereby allowing them to scrape large amounts of Reddit data from search results.

It is said that the purpose of these tools is to bypass two layers of protection: first, to evade Reddit’s own anti-crawling systems; second, to bypass Google’s controls and directly extract Reddit content from its search results.

The lawsuit states that these data companies operate as “data crawling service providers,” allowing them to access nearly 3 billion search result pages without authorization, thereby circumventing Google’s technical controls.

Reddit claims that even after receiving a cease-and-desist letter in May 2024, Perplexity continued to use the data from these three companies in its answer engine.

A representative of Perplexity responded to this by sharing a full reply on Reddit.

Perplexity deliberately posted its response on Reddit, aiming to convey the idea that it was a publicly accessible link to Reddit. But according to Reddit’s lawsuit, if anyone quotes this link in any way, they could also be sued. The representative told Decrypt this.

Perplexity described this lawsuit as “a tragic example of how public data becomes an important part of a company’s business model.”

“Reddit believes this is their right. But this goes against the principles of an open internet,” said Perplexity.

A representative of SerpApi told Decrypt that they have not received any communication or services from Reddit regarding this matter. They added that they strongly reject Reddit’s accusations and plan to take legal action.

Denas Grybauskas, the Chief Governance and Strategy Officer at Oxylabs, stated in an email statement to Decrypt: “No company should claim to own public data that does not belong to them. Such claims may simply involve selling that same public data at high prices.” Grybauskas also mentioned that Reddit “did not attempt to communicate with Oxylabs at all.” Decrypt has contacted Reddit, Google, and AWM Proxy to seek their opinions. If they respond, the article will be updated accordingly. Andrew Rossow, a legal expert specializing in public affairs and the director of strategic partnerships for video search and content intelligence platforms like Oriane, told Decrypt that in such cases, courts need to determine whether the terms of service provided by platforms like Reddit explicitly include provisions related to artificial intelligence training, data scraping, and commercial use. Rossow explained that if users agree to grant the platform broad, permanent, royalty-free permissions to their content, those permissions typically govern the relationship between users and the platform. However, he added that artificial intelligence companies are not automatically granted such permissions unless the terms allow the platform to re-license or sell data for such purposes. Rossow emphasized that courts must distinguish between users’ rights to their expressed opinions (post text) and the use of that content for data mining purposes (such as extracting patterns, facts, and language models). Despite this, Rossow believes that the so-called “knowledge” behind LLMs is the result of millions of users’ time, effort, and creative expression.

Rossow believes that treating content created by humans as free, unmodified resources is a form of labor exploitation, and it diminishes the value of online contributions. He also said that AI companies need to “respect the norms and rules of digital citizenship and communities,” as these are the implicit and explicit rules that govern the digital public space they operate in.