Meta's spokesperson Andy Stone today called reports that the company's updated commercial terms would allow it to scrape news sites "false," a day after Nieman Lab published the views of three legal experts who said the text, as written, could support exactly that reading. The revised Meta Commercial Terms take effect on October 30, 2026, and contain a new paragraph in which every business using Meta's products authorises Meta to send "web scrapers, bots, robots, spiders, crawlers and user agents" to the websites associated with its products or services, and to use the content collected there to "train" Meta Company Products.

In Short

Meta rewrote the rules that every business on Facebook and Instagram agrees to, and the new version says Meta's robots can visit those businesses' websites, copy what they find and use it to build and train Meta's products, including AI. News publishers, which almost all run Meta business accounts, worry this lets Meta take their articles for free, while Meta says the change is only about reading product details from shopping sites. The rules start on October 30, 2026, and the only way to reject them is to stop using Meta's business products.

What the October 30 terms add

The clause sits inside Section 1 of the new terms, which Meta has renamed from "Licences" to "Licences and rights you provide Meta." It is the most consequential addition in a document that otherwise carries over most of the March 6, 2024 version word for word. Here is the full text, as it appears in the UK English edition of the terms dated October 30, 2026:

"You authorise us (or any entity acting on our behalf) to use automated software (such as web scrapers, bots, robots, spiders, crawlers and user agents) and create test credentials to access, index, cache, analyse, collect, retrieve or crawl the domains, web pages or URLs (and any content available through them) that are associated with your products or services, or that are made available through your or users' use of Meta Company Products and services, and to periodically pull updates from such domains, web pages or URLs. 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."

Nothing resembling that paragraph exists in the version effective March 6, 2024. The old Section 1 ran to four sentences and dealt only with content that a business "share[s], post[s] or upload[s] on or in connection with" Meta's products. It required the business to secure rights for Meta "to display, distribute and deliver your content within the Meta Products." The new section widens that list. Businesses now represent that they hold rights permitting Meta to "display, distribute, create derivative works and deliver your Content," and a new sentence states that Meta "may use such Content as described in our Terms, including to provide, develop, improve, research and maintain Meta Company Products, services and features."

The crawling paragraph goes further than either of those changes because it reaches content that never touches Meta's platforms. The verbs are broad: access, index, cache, analyse, collect, retrieve and crawl. So is the scope - "any content available through" the domains, pages or URLs in question. Meta also reserves the right to "periodically pull updates," which describes a recurring crawl rather than a one-time fetch.

Two phrases carry most of the weight. The first is "associated with your products or services." The terms do not define it. A publisher whose product is journalism could read its entire site as associated with its products; a retailer could read it as covering only product pages. The second phrase is "made available through your or users' use of Meta Company Products and services." That prong is not tied to the business's own domains on its face. A URL becomes covered when it is made available through a business's use of Meta's products - or through users' use of them. The text does not say how a business could grant permission for sites it does not operate, and none of the documents reviewed for this article address that question.

The third element is "test credentials." The clause lets Meta, or any entity acting for it, create them in order to access the covered domains. The terms do not explain what test credentials are or where they would be used.

Who is bound, and how

The terms apply to "access or use of the Meta Company Products (or 'Products'), for a business or commercial purpose." The new definition of business purposes is longer than the old one. The March 2024 text listed "using ads, selling products, developing apps, managing a Page, managing a Group for business purposes or using our measurement services." The October 30 version adds "marketing messages," "listing products," "creating or using a product catalogue," and "using any business product, such as Ads Manager or Meta Business Suite."

Consent is passive. Section 7 of the new terms, which replaces the old Section 6, states: "By continuing any access or use of any Meta Company Products for business or commercial purposes after any notice of an update to these Commercial Terms, you agree to be bound by them." A business that posts to its Facebook Page on October 31 has, under that wording, accepted the crawling clause.

The new terms also name the Instagram Terms of Use alongside the Meta Terms of Service as applying to commercial use "except where we state otherwise." The March 2024 version referred only to the Meta Terms of Service. Neither version of the Commercial Terms reviewed here mentions WhatsApp by name, although Nieman Lab's reporting describes the change as affecting accounts that manage "WhatsApp channels" and Jason Kint's post lists WhatsApp among the covered services. The phrase "Meta Company Products" is not defined in the Commercial Terms themselves.

A wording change Nieman Lab quoted from the old text

According to Nieman Lab, there is currently no option for business accounts to opt out of the new terms ahead of October 30. The publication quoted the update as reading: "If you do not agree to the updated terms, please stop all access or use of our Products for business or commercial purposes." That sentence is the March 2024 wording. The October 30 version softens it to: "If you do not agree to the updated terms, you may stop all access or use of our Products for business or commercial purposes." The practical effect is the same. Rejecting the terms still means leaving Meta's commercial products, and neither version offers a partial opt-out from the crawling paragraph.

Meta's position: product data, not news

Meta has now offered two accounts of the clause. The first came through Nieman Lab. Meta declined to answer the publication's list of questions, but a public affairs spokesperson said in an email that language around crawling had previously been scattered across different terms of use, "including terms about using generative AI in ads," and that the update consolidates it in one place. According to Nieman Lab, the spokesperson also said the new terms will not change how Meta crawls publishers in the future and that "content owners can still manage how their content is accessed by using robots.txt."

The old Commercial Terms contain no crawling language at all. That is consistent with the claim that the wording previously lived elsewhere - in supplemental or product-specific terms that were not part of the documents supplied for this article. It does not confirm it either.

The second account came today, on X. Andy Stone, whose profile lists him in communications at Meta, responded to a Techmeme post summarising the Nieman Lab story. "This is false," Stone wrote. "There's no change to how Meta uses web crawlers. As always, content owners can use robots.txt flags to indicate their preferences. This update is designed to make sure we're able to pull accurate information about products for sale on third-party websites when making on-platform ad and content recommendations, not about scraping news sites." The post displayed 3,679 views at the time it was captured.

Stone's explanation fits part of the text. "Associated with your products or services" reads naturally as a reference to product pages, and Meta has spent October building shopping features that depend on accurate product data. PPC Land reported on October 15 that Meta will turn on product browsing in existing ads on October 28, two days before the new terms take effect, so a tap on an ad may open an in-app page with product information before the advertiser's site. Meta's Muse agent, which gained access to ad accounts and 15 business tools when it expanded to small businesses on September 29, also relies on product information that lives outside Meta's apps.

But the paragraph itself contains no product limitation. It does not mention products for sale, catalogues, prices or recommendations. The permitted purposes are "provide, develop, improve, research, train and maintain the integrity of" Meta Company Products, services and features. Stone's statement describes an intent; the contract describes a permission. Those can be different things, and the contract is what binds the parties after October 30.

Nieman Lab staff writers Andrew Deck and Hanaa' Tameez published their report on October 15 at 12:59 p.m. Three legal experts reviewed the terms at the publication's request. Their comments are interpretations of the contract text, not findings by any court or regulator, and in places they describe the outer limit of what the words could support rather than what Meta has done.

Matt Topic, a media and intellectual property attorney at Loevy & Loevy, focused on value. "This is a remarkable provision to bury in [the] terms of service, given that this would give Meta access to material that would likely cost many millions of dollars in a licensing market," Topic said, according to Nieman Lab. "It's not clear whether a court would enforce this, but publishers should assume that it would be enforceable."

James Grimmelmann, a professor of digital and information law at Cornell Law School, read the clause as a training licence. "For news publishers, this clearly represents 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," he said. He then described the widest reading available. "On the broadest possible interpretation, this would be used to develop a comprehensive archive of all content that Meta would then serve as it wanted to directly or through AI." The word "train" does appear in the clause; the "comprehensive archive" is Grimmelmann's description of a possible use, not a term in the contract.

Grimmelmann also addressed test credentials. According to Nieman Lab, he said this would technically allow Meta to "sign up as a fake user" to get past registration walls on a publisher's site. Meta declined to tell Nieman Lab whether test credentials would be used to bypass paywalls or other paid content. The terms say only that Meta may "create test credentials to access" covered domains; they do not mention registration walls or paywalls.

On choice, Grimmelmann was blunt: "publishers either have to trust that Meta doesn't mean the full extent of what these terms say, or they have to stop [using business accounts]."

Benjamin Sobel, a law professor at the University of Wisconsin, framed the change in antitrust terms. Tethering access to business products to scraping permissions echoes a concept called "tying," according to Sobel. "If you want to use the Meta products that require acceptance of the commercial terms, you have to also acquiesce to the scraping and crawling permissions exacted in those terms," he said. Sobel compared it with Google's approach to AI search, under which publishers must accept use in AI Overviews and AI Mode to appear in traditional results. No tying claim has been filed over the terms, according to the documents reviewed.

The legal experts also told Nieman Lab that the new language creates an opening for Meta to argue that its commercial terms supersede other blocking instructions. That is the point where Stone's robots.txt reassurance and the contract meet. The terms do not mention robots.txt. They neither carve it out nor override it.

Jason Kint's thread

Jason Kint, chief executive of Digital Content Next, a trade group for news and entertainment publishers, posted a five-part thread on X early on October 16. The first post displayed 45,500 views at the time of capture.

"This reads that Meta is looking to grant itself the rights to scrape websites across the open web - and train Meta's AI - once a company uses Facebook, Instagram, WhatsApp or other Meta services after October 30, 2026," Kint wrote. "So every biz."

The phrase "across the open web" goes beyond the clause's literal wording, which limits the first prong to sites "associated with your products or services." The second prong, covering URLs "made available through your or users' use" of Meta's products, is broader and does not specify whose sites those are, which is the textual basis for wider readings. Kint's second post narrowed the claim to the clause's own words: "Its new terms reach outside its walled gardens to websites 'associated with your products.'" He suggested the content could power Meta's AI, "including Muse, without licensing deals" - an inference about future use, not a statement in the terms.

In his third post, Kint addressed Meta's consolidation explanation: "Then why demand sweeping permissions in a binding contract? Why explicitly authorize AI training? Words matter." The fourth tied the clause to litigation over books, saying Meta "faces publishers' lawsuits alleging it torrented pirated books from shadow libraries to train its AI" and that a trial had been "set just yesterday for May 3." That trial date is Kint's account and could not be confirmed from the supplied documents. Nieman Lab separately reported a class action filed in May by five large book publishers alleging Meta pirated millions of books to train its Llama models. The fifth post pointed to a ProPublica article on Facebook's tracking history and concluded: "Now Meta says these terms won't change its behavior. We're supposed to trust that over the words it wrote into terms? No."

Kint then replied to Stone's denial directly. "The terms are being changed to allow it," he wrote. "That you're not doing it now or the ietf signal shouldn't comfort any website." The reference to an IETF signal concerns standards work on machine-readable AI preferences; Kint did not elaborate in the captured reply.

What else changed in the terms

The crawling clause drew the attention, but the October 30 rewrite touches almost every section. Several changes shift risk toward businesses.

New AI features section. Section 5 is entirely new. It states that Meta "may use or make available products and features powered by artificial intelligence, machine learning, generative AI or other automated products or features." Businesses "acknowledge and agree that AI features may make mistakes and may not always produce accurate, complete or suitable results" and are "solely responsible for and assume any and all risks associated with such AI features and any resulting Content." The section also prohibits altering or removing "any watermark or content-authenticating metadata generated by AI in connection with Content." A further sentence makes businesses "solely responsible for the acts or omissions of any third-party software or service" they use or authorise with Meta's products.

Warranty disclaimer. Section 4, renamed "No Warranties; Limits on liability," now states that Meta's products "are provided 'AS IS'" and disclaims "ALL WARRANTIES, WHETHER EXPRESS OR IMPLIED, INCLUDING THE IMPLIED WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE, TITLE AND NON-INFRINGEMENT." The liability cap is unchanged at the greater of $100 or the amount paid to Meta in the previous twelve months, now prefaced with "To the extent permitted by law."

Wider indemnity. The old clause, headed "Third-party claims," began "If anyone brings a claim." The new clause, headed "Indemnification," begins "If you or anyone brings a claim." That small addition extends the indemnity to claims brought by the business itself.

Supplemental terms folded in. A new Section 8 incorporates five sets of terms by reference: the Self-serve Ad Terms, the Business Tools Terms (covering Meta Pixel, Conversions API, the Facebook SDK for App Events, Offline Conversions and the App Events API), the Seller Agreement, the Meta Business Agents and Platform Terms of Service, and the Meta AI Business Assistant Supplemental Terms. The Business Agents entry defines its scope as including "Meta Business Agent Connectors and assistants from Meta and other software, tools, services, documentation, content, functionalities and generative AI features."

Discretion to discontinue. A new clause 9(g) states that Meta "reserves the right, at its sole discretion, to modify, limit, suspend or discontinue any commercial products, services and features at any time without prior notice or liability."

Severability. The old text severed unenforceable portions. The new one first deems them "amended to the minimum extent necessary to make it enforceable," and severs them only if that fails.

Address. The notice address for arbitration moves from "1 Meta Way, Menlo Park" to "1601 Willow Rd. Menlo Park."

The dispute section, now Section 6, is otherwise carried over. One unchanged sentence sits awkwardly beside the new crawling clause. Arbitration for US businesses excludes claims relating to "efforts to interfere with our Products or engage with our Products in unauthorised ways (for example, automated ways)," and Section 6.d.iv lets Meta sue businesses outside the United States in their home country over "efforts to abuse, interfere or engage with our Products in unauthorised ways." Meta's terms treat automated access to its own products as a matter it reserves the right to litigate, while the new Section 1 grants Meta automated access to its business users' sites.

Why it matters for publishers and advertisers

For advertisers, the most direct effect may be on product data. If Stone's description is accurate, Meta wants contractual cover to read product pages on advertisers' sites so its ad and content recommendations reflect current prices and stock. That would align with the product browsing change on October 28 and with Muse's shopping functions. Advertisers that already submit catalogue feeds would see little practical difference. The legal grant, however, is not limited to product pages or to advertisers, and it applies equally to businesses that never run an ad but manage a Page.

For publishers, the clause lands in a market where crawler controls are already contested. Meta runs several crawlers with different jobs. Meta-ExternalAgent collects public pages both to train foundation models and to index content for Meta's products, using a single robots.txt token for both purposes. That makes it a mixed-use crawler: a publisher cannot allow indexing while refusing training. A separate agent, Meta-WebIndexer, builds the index behind Meta AI answers, and FacebookExternalHit fetches pages only when a link is shared to generate a preview card.

Robots.txt compliance is voluntary. That is the gap the legal experts identified: Meta's public documentation says it honours robots.txt for its crawlers, but a contract signed by the site owner could, in principle, be argued to override a technical signal. Meta has said it will not use the terms that way. The terms do not say so.

The timing also matters. Meta's crawling footprint has grown quickly. PPC Land reported in September 2025 that Meta's share of AI crawling traffic rose from 0.9% to 7.5% between July 2024 and July 2025. On October 15, DataDome found that Meta's Muse accounted for 33% of agentic browser traffic in September, with 99.3% of a sampled set of its requests coming from cloud and CDN infrastructure.

The broader legal backdrop is unsettled. Any attempt to train Meta's AI on content collected under the new terms would still depend on open questions about AI training and fair use, according to the experts Nieman Lab consulted. In June 2025, Judge Vince Chhabria granted Meta summary judgment in Kadrey v. Meta, finding its use of books from thirteen authors to train Llama was fair use, though the ruling applied only to those plaintiffs. Two days earlier, Judge William Alsup had found fair use in AI training but rejected Anthropic's piracy defense. A contractual licence would make fair use less relevant for any site covered by it - which is part of why the wording draws scrutiny.

Other companies have faced similar complaints about bundling. Publishers have argued for more than a year that Google offers no way to opt out of AI Overviews without losing visibility in search, the comparison Sobel drew. In the UK, Google added a Search Console toggle for AI features on June 3, 2026, the same day the Competition and Markets Authority imposed its first conduct requirement. In September, Cloudflare dropped a planned Googlebot block and introduced a "Disallow AI Training" option instead. Legislators have also moved: New York's Stealth Crawler Prohibition Act would require crawlers on news sites to identify themselves and disclose their purpose, with penalties of up to $15,000 a day.

Meta's own history with training data is relevant context too. In April 2025, Meta said it would use public posts and comments from adult users in Europe to train its AI models from May 27 that year, and offered an objection form. The new commercial terms offer no equivalent form for the crawling paragraph.

Nieman Lab also noted a commercial tension. Meta has begun restricting the number of external links some Facebook and Instagram accounts can post without a paid subscription, which the publication called Meta One, while the new terms let it crawl linked content without charge. Nieman Lab cited Press Gazette as reporting that some news publishers are currently exempt from the link limit. Meta declined to answer Nieman Lab's question about the service.

One small business owner gave the story its origin. Stephen Beale, founder and editor of The Steampunk Explorer, saw the notice in his Meta Business Suite dashboard and tipped off Nieman Lab. "But this was different because they're saying that [since you are] a business user of Facebook, they have the right to go out and grab information from basically wherever you have a presence," Beale said. He told the publication he is seriously considering shutting down his accounts before October 30. "I know my site has been scraped to death," he said. "But it just seems like such an arrogant, in-your-face move on [Meta's] part."

What remains open is whether Meta will narrow the text before it takes effect. Stone's denial and the spokesperson's statement to Nieman Lab both say Meta's crawling will not change. Neither says the clause will be rewritten. Fourteen days separate today's exchange from the date on which continued use of a Facebook Page, an ad account or Meta Business Suite becomes acceptance.

Timeline

Summary

Who: Meta, through its updated Commercial Terms; Meta communications spokesperson Andy Stone; Nieman Lab reporters Andrew Deck and Hanaa' Tameez and the legal experts they consulted (Matt Topic of Loevy & Loevy, James Grimmelmann of Cornell Law School, Benjamin Sobel of the University of Wisconsin); Digital Content Next chief executive Jason Kint; and every business using Meta's products for commercial purposes, including news publishers and advertisers.

What: A new paragraph in Section 1 of the Meta Commercial Terms authorises Meta to use automated software and test credentials to crawl domains associated with a business's products or services, or made available through business or user activity on Meta's products, and to use that content to "train" and develop Meta products. Legal experts told Nieman Lab the text could support scraping news sites for AI training; Meta says the update concerns product information and does not change its crawling. The rewrite also adds an AI features section, a warranty disclaimer, a wider indemnity and incorporated supplemental terms.

When: The terms take effect on October 30, 2026, replacing the version effective March 6, 2024. Nieman Lab reported on October 15, 2026; Kint and Stone posted on October 16, 2026.

Where: The terms apply to commercial use of Meta Company Products worldwide, with US businesses bound to individual arbitration and non-US disputes routed to courts in California or Ireland.

Why: Consent comes from continued use, there is no opt-out short of leaving Meta's business products, and the clause names AI training without limiting its scope to product pages - a gap between Meta's stated intent and the contract's wording that publishers and advertisers have fourteen days to weigh.