This Wednesday, online forum Reddit filed a lawsuit in San Francisco Superior Court against artificial intelligence company Anthropic, accusing it of breaching contractual terms, circumventing technical barriers, and scraping user-generated content.

According to the complaint, Anthropic’s use of web crawlers to collect Reddit content violates Reddit’s user agreement and constitutes unfair competition under California law. Reddit alleges that Anthropic deliberately used this data to train its AI models without obtaining user consent.

The lawsuit further claims that Anthropic’s 2024 public statements—asserting that it had limited its content-scraping bots following complaints—were insincere. The complaint references a July 2024 report by repair community website iFixit, which stated that Anthropic’s crawler accessed its site more than one million times in a single day. At the time, major AI companies, including Anthropic, publicly claimed that their crawlers respected access restrictions set via websites’ robots.txt files. While compliance with robots.txt is not legally mandatory, websites can cite violations as evidence of improper conduct. Reddit argues that Anthropic failed to adhere to such restrictions, though it does not detail specific violations occurring after July 2024.

Reddit emphasized that it has entered into content licensing agreements with several companies, including OpenAI, Google, Sprinklr, and Cision. These agreements allow AI companies to legally use Reddit content for model training while respecting users' rights to delete their posts. However, Reddit states that Anthropic refused to engage in any such licensing negotiations.

In the complaint, Reddit claims it detected at least 100,000 instances in which Anthropic’s automated tools accessed—or attempted to access—Reddit content in clear violation of Reddit’s terms and despite multiple cease requests. Reddit argues that this conduct represents a continuous attempt by Anthropic to extract value from Reddit while disregarding legal and ethical boundaries.

A Reddit spokesperson stated that the company believes in an open internet but that this “does not grant Anthropic the right to illegally scrape Reddit’s content, profit in the billions, and ignore users’ rights and privacy.”

Anthropic has denied Reddit’s allegations. A company spokesperson said, “We disagree with Reddit’s claims and will vigorously defend ourselves.”

The lawsuit comes at a time when AI companies face soaring demand for online data, while website owners are increasingly concerned about how their content is being used to train models—and what compensation and user privacy protections are in place. Reddit signed a content licensing agreement with OpenAI in May 2024. In its Q1 2025 earnings call, CEO Steve Huffman emphasized the importance of a steady supply of fresh information to search and AI companies, framing content licensing as a core part of Reddit’s business strategy.