Meta's new commercial terms could let it scrape news sites to train AI, legal experts warn

What happened

Nieman Lab reported on October 16, 2026 that Meta's updated Commercial Terms of Use, set to take effect October 30, authorize Meta to use automated software (web scrapers, bots, crawlers and similar tools) to access, index, cache, analyze, collect and crawl domains, web pages or URLs "associated with" business accounts on Facebook, Instagram and WhatsApp — and to use any collected content to "provide, develop, improve, research, train, and maintain the integrity of" Meta's products.

The legal reading

Three legal experts who reviewed the terms told Nieman Lab the language could let Meta argue it has contractual permission to scrape news publishers' own websites and use that content to train AI products including Llama, without a separate license. Matt Topic (Loevy & Loevy) called it "a remarkable provision to bury in [the] terms of service" and said publishers should assume it would be enforceable; James Grimmelmann (Cornell) said that on the broadest reading Meta could use it "to develop a comprehensive archive of all content" and deploy it in competition with publishers. The experts also noted the "test credentials" clause could let Meta bypass registration walls. There is no opt-out: business accounts must stop using Meta's commercial products to avoid the terms.

Meta's response

Meta declined to answer Nieman Lab's questions but said the language consolidated crawling provisions previously scattered across its terms, that crawling behavior is unchanged, and that content owners can manage access via robots.txt.

Status

The experts' interpretation is analysis, not a court ruling; whether courts would enforce the terms against publishers is unsettled, and training-use questions remain subject to ongoing litigation (including a class action by book publishers alleging Meta pirated books to train Llama).