The checklist of lawsuits towards AI firms is rising: Reddit has joined in with a go well with towards Anthropic.
On Wednesday, the corporate filed a criticism in California stating that Anthropic — developer of Claude — ignores Robots Exclusion Protocol (REP), or robots.txt, which blocks AI crawlers from scraping a web site’s content material. Analysis signifies that different AI firms are additionally partaking on this apply: In March, Columbia’s Tow Heart discovered that a number of chatbots, together with Perplexity, might nonetheless retrieve articles from publishers that had blocked their crawlers utilizing REP.
The criticism states that “Anthropic is the truth is deliberately skilled on the private knowledge of Reddit customers with out ever requesting their consent,” which is a violation of Reddit’s person privateness settlement. In July 2024, when Reddit publicly criticized Anthropic for misusing its content material, the criticism continues, “Anthropic’s bots continued to hit Reddit’s servers over 100,000 instances” regardless of insisting that it had stopped its bots from crawling the positioning.
The lawsuit is the most recent within the ongoing conflict between websites that create and host content material — together with publishers, information organizations, and person boards like Reddit — and the AI firms that scrape that content material to make use of as coaching knowledge. In late 2023, The New York Occasions turned the primary writer to sue OpenAI and Microsoft for utilizing its content material to coach its fashions with out permission or fee. In April, Ziff Davis, the guardian firm of this publication, sued OpenAI for copyright violation, citing comparable cases of the AI firm crawling Ziff Davis websites regardless of being blocked. Authors and creatives have additionally sued OpenAI and Meta on comparable grounds.
What units Reddit aside right here is that it’s also a tech firm, not like the publishers behind the lawsuits that predate this one. Reddit has licensing agreements with OpenAI and Google.
Different publishers, together with Dotdash Meredith, Monetary Occasions, and the AP, have taken a special strategy, proactively getting into into licensing agreements with AI firms that permit them to entry some or all of their content material in change for inside AI instruments and preferential quotation placements in chatbot responses. Nonetheless, analysis exhibits that chatbots nonetheless battle to precisely cite and favor tales from publishers, which means it’s nonetheless unclear whether or not these advantages are being realized.
Need extra tales about AI? Join Innovation, our weekly publication.
