Meta Strikes AI-Training Clause From Commercial Terms After Legal Pushback

Meta has quietly rewritten its commercial terms page to delete a clause that gave the company broad rights to scrape and reuse content from across the web, including news sites, to build its AI systems, The Verge reported.

The deleted text stated that "Meta may use any content from such domains, web pages, or URLs to provide, develop, improve, research, train, and maintain the integrity of, the Meta Company Products, services, and features provided under these Commercial Terms and any other applicable supplemental terms." That single sentence functioned as a standing license: any site operator who agreed to Meta's commercial terms for an unrelated product, such as advertising or business tools, could be read as simultaneously granting Meta permission to feed its content into AI training pipelines.

The clause drew scrutiny from Nieman Lab and from James Grimmelmann, a professor of digital and information law at Cornell, who read it as sweeping enough to cover news publishers that had never negotiated an AI licensing deal with Meta. Unlike the individual content-licensing agreements Meta has struck with specific publishers, this language sat inside the general commercial terms that countless businesses accept to use Meta's advertising and marketing products, making it a default grant rather than a bargained-for exchange.

The Verge noted this is not the first time Meta has edited this particular page under pressure. The company has revised the commercial terms before in response to criticism, and the latest change follows the same pattern: quietly remove the disputed clause rather than publicly litigate its meaning.

With the sentence gone, Meta no longer has contractual text on that page it can point to as authorizing scraping of a counterparty's domains, web pages or URLs for AI training. Site operators who previously accepted the commercial terms are left without a clear record of what, if anything, Meta now claims it can still do with content they published while the broader clause was in effect.

Meta did not respond to a request for comment from The Verge about why the language was removed or whether it affects content already used to train existing models.