Meta's New Commercial Terms Could Let It Scrape News Sites to Train AI

Meta's New Commercial Terms Turn a Business Account Into a Free AI Scraping Licence

Meta told business users in mid-October, through a help blog and the Meta Business Suite console, that new Commercial Terms take effect October 30. After reading them, three legal experts and a growing list of publishers say the fine print hands Meta the right to crawl their websites — and with it, the ability to skip millions of dollars in AI licensing fees.

October 30, and no way to say no

The update covers commercial accounts on Facebook, Instagram, WhatsApp and Messenger. Meta defines "commercial" broadly: advertising, marketing, selling or listing products, developing apps, keeping a product catalogue, running a Page or business Group, or using a business product such as Meta Business Suite. Virtually every news organisation holds at least one such account. The new terms replace a version dated March 6, 2024. There is no opt-out: businesses that disagree are told to "stop all access or use of our Products for business or commercial purposes."

What the clause actually authorises

Under the heading "License & Rights You Provide," a business authorises Meta, "or any entity acting on our behalf," to use automated software — scrapers, bots, robots, spiders, crawlers, user agents — and to "create test credentials" in order to access, index, cache, analyse, collect, retrieve or crawl domains, web pages and URLs associated with its products or services, including links surfacing through the business's or users' use of Meta products, and to pull periodic updates. Meta may use that content to "provide, develop, improve, research, train, and maintain the integrity of" its products. The word "AI" never appears in the clause; elsewhere the terms list "AI at Meta" among those products.

Meta says crawler behaviour is unchanged

Spokesperson Daniel Roberts told reporters the company is simply "consolidating" its commercial and self-serve ad terms, adding that "there is no change to how Meta uses web crawlers; any suggestion to the contrary is false," and that content owners can still use robots.txt. Meta's Andy Stone called the coverage false on X, writing that the point is pulling accurate information about products for sale on third-party sites "not about scraping news sites." Meta also told Nieman Lab it has no plans to use the terms to ignore robots.txt on news sites.

Experts: a licence worth millions, buried in fine print

Matt Topic, a media and intellectual property attorney at Loevy & Loevy, called it "a remarkable provision to bury in terms of service," because it gives Meta access to material that "would likely cost many millions of dollars in a licensing market." That is a market estimate; Meta has published no savings figure. Cornell law professor James Grimmelmann said the terms amount to "permission to Meta to train AI on any content from their sites and possibly to deploy some of that content, even in competition with the publishers themselves." The three experts also warned the wording lets Meta argue its commercial terms override other blocking instructions, and likened the arrangement to antitrust "tying" — much as Google requires sites to allow AI Overviews scraping to stay in search. "Test credentials" is new; Grimmelmann said it would technically let Meta "sign up as a fake user" to clear registration walls. Meta declined to say whether it would also reach paywalled content.

Publishers already sell what Meta would take free

Meta does pay for some content: CNN, Fox News, USA Today, People Inc. and Le Monde Group are among its licensing partners, and its News Corp deal is reported at up to $50 million a year for at least three years. Market benchmarks are public: News Corp and OpenAI at $250 million over five years, Amazon's New York Times licence reported at $20–25 million a year, the Financial Times and OpenAI at $5–10 million a year. A 2026 tally of 312 licensing deals found news publishers took 104 of them, at a median $4.2 million a year. Under the new terms, none of that negotiation is needed — staying on Meta's platforms counts as consent.

Charged to post links, allowed to crawl the content free

Meta One, launched September 15, puts clickable links in organic Instagram posts and Reels behind the $49.99-a-month Advanced tier, which allows four linked posts and four linked Reels a month; Expert costs $149.99 and Max $499.99, while free Facebook Pages get two linked posts or comments monthly. As Nieman Lab put it, Meta has begun charging users to post links to the same content it will be permitted to crawl for free. The News/Media Alliance's Danielle Coffey called the terms "outrageous and probably illegal," asking whether Meta would claim access to copyrighted material on Taylor Swift's website if she holds a business account. Angela Mills Wade of the European Publishers Council said holding a business account, or having a reader share a link, "should never be treated as a blanket licence."

A bill in Congress, a lawsuit in California

Before October 30, businesses get no transition period and no regional or size-based carve-out. In Washington, the bipartisan Stealth Bot Prohibition Act (H.R. 9915), introduced July 23, would require crawlers to disclose their identity and purpose, with Federal Trade Commission civil penalties; it has not passed, though New York enacted a similar state law in June and more than 300 publishers lobbied Congress on September 29. In court, a suit by five publishers and writer Scott Turow over Meta's AI training was ordered transferred from Manhattan to Northern California on September 25, joining six similar cases already in California. And technically, Meta's own documentation lists its training crawler as Meta-ExternalAgent, blockable through robots.txt; a September 24 survey of 1,065 robots.txt files found 117 naming it and 51 blocking it at the root.