Meta just changed the rules of the game, and most players don’t even know they’re on the field.
Buried in an updated set of commercial terms that takes effect October 30, 2026, Meta now claims the right to crawl and scrape websites “associated” with any business account on Facebook, Instagram, WhatsApp, or Messenger — and to use that content to “develop, improve, research, and train” its AI products. Legal experts who reviewed the language say this effectively gives Meta permission to train AI on news sites’ content for free, without a licensing deal, and potentially without regard for robots.txt blocking instructions.
The new policy applies to the roughly 200 million business accounts across Meta’s platforms. That includes virtually every news publisher, e-commerce brand, and content creator that maintains a presence on Facebook or Instagram. If you promote your products or services through Meta’s business tools, you’ve already accepted these terms — or you will by October 30.
The Language That Launched a Thousand Alarms
The critical clause sits under “License & Rights You Provide” in Meta’s commercial terms. It authorizes Meta to “use automated software (such as web scrapers, bots, robots, spiders, crawlers, and user agents) and create test credentials to access, index, cache, analyze, collect, retrieve, or crawl the domains, web pages, or URLs” associated with your business.
James Grimmelmann, a professor of digital and information law at Cornell University, reviewed the terms for Nieman Lab. His assessment was blunt: “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.”
That last part is the kicker. It’s not just about training a model in the background. The language suggests Meta could deploy scraped news content directly in its AI products — competing with the very publishers whose work it extracted.
Meta Pushes Back — But the Terms Say What They Say
Meta spokesperson Daniel Roberts dismissed the reporting. “There is no change to how Meta uses web crawlers; any suggestion to the contrary is false,” Roberts told Nieman Lab. “Content owners can still use robots.txt flags to indicate their preferences.”
Andy Stone, Meta’s vice president of communications, echoed that on X: “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.”
But here’s where it gets legally interesting. Grimmelmann and other experts tell Nieman Lab that the commercial terms could be interpreted as superseding other blocking instructions — including robots.txt files. If Meta’s position is that its commercial terms grant it access, then a publisher’s robots.txt file might not be enough to keep Meta’s crawlers out. The terms create a contractual pathway that didn’t exist before.
Whether Meta would actually exercise that pathway is another question. The company’s public stance is that it respects robots.txt and has no plans to use the terms to bypass publisher preferences. But the fact that the language is there — in a binding legal document that businesses must accept by October 30 — changes the leverage dynamics.
Meta’s Existing Licensing Deals: The Big-Player Exception
Meta has been busy signing AI licensing deals with major publishers. In March 2026, News Corp struck a multi-year deal worth up to $50 million per year, letting Meta use content from the Wall Street Journal, The Post, and other News Corp properties. That followed earlier deals with CNN, Fox News, USA Today, and People Inc. in December 2025. This is the same company that just launched Meta’s Muse Spark Open 816B model — a frontier AI system that needs vast amounts of training data.
These deals give Meta’s AI assistant access to real-time news content. They also give publishers a share of the AI revenue pie. But they cover only the largest players. The new commercial terms would extend Meta’s reach far beyond these deals — to every business account holder, from the smallest local blog to mid-sized online publications that never had the leverage to negotiate a licensing agreement.
Put another way: Meta has created a two-tier system. The big publishers get paid. Everyone else’s content is available for free under terms they didn’t individually negotiate and can’t opt out of without leaving Meta’s platforms entirely.
The Coercion Pattern: Accept or Leave
Businesses that don’t agree to the updated terms “may stop all access or use of our Products for business or commercial purposes.” That’s not an opt-out. That’s an ultimatum. Stay on Meta’s platforms and accept the new terms, or leave.
For news publishers, leaving Meta isn’t a realistic option. Meta’s platforms are where audiences discover content. A publisher that walks away from Facebook and Instagram loses reach, referrals, and ad revenue. The terms are structured so that the cost of saying no is higher than the cost of saying yes — which is exactly why legal experts are alarmed. It’s the same playbook as Samsung’s consent-or-delete health data policy: take it or leave it.
This pattern should feel familiar. We’ve seen it before with platforms that extract value from creators and then present a take-it-or-leave-it choice. Twitch did it with AI training defaults. Samsung did it with health data. Now Meta is doing it with news content.
The Broader Context: Who Owns the Web’s Content?
Meta’s move comes amid a broader fight over who owns the content that trains AI models. OpenAI has signed licensing deals with the Wall Street Journal, Financial Times, and Vox Media. And the AI industry’s trust problem keeps growing — from security blind spots inside OpenAI to platform-level data extraction. Google says more than 400 publications participate in its news AI pilot program. But at the same time, AI companies face lawsuits from publishers who say their content was scraped without permission.
The Seattle Times and Newsday sued OpenAI and Microsoft in September 2026, alleging they copied copyrighted journalism without permission or payment. The New York Times has a case against OpenAI that’s been grinding through courts since 2023, part of the broader AI training data copyright fight. Publishers are both licensing their content and suing over the same practice — a reflection of how unsettled the legal landscape is.
Against this backdrop, Meta’s commercial terms look less like an isolated policy update and more like a power play. If the terms hold up in court, Meta gains a legal framework for scraping content that other companies lack. If they don’t, Meta faces the same legal risks its peers are confronting. Either way, the people whose content is being extracted — the publishers, the journalists, the creators — have the least say in the matter.
What This Means for Creators and Small Publishers
For a major publisher like News Corp, the calculus is clear: negotiate a licensing deal and get paid. For a mid-sized online publication or an independent journalist with a Facebook page, there’s no such option. The terms apply to them whether they like it or not.
Consider the practical scenario. A regional news outlet maintains a Facebook page to reach readers. It posts links to articles, which drive traffic to its website. Under the new terms, Meta can crawl that outlet’s website, extract its content, and use it to train AI models — without paying the outlet anything. The outlet’s only choices are to accept this or leave Facebook, which would mean losing its primary distribution channel.
That’s not consent. That’s coercion dressed up as a terms-of-service update.
And it’s not just news. Any business with a website and a Meta presence — from a local restaurant to an online store to a personal blog — falls under these terms. The scope is enormous.
The Philippines Angle: A Cautionary Tale for Regulators
Here in the Philippines, we’re no strangers to platform power. Facebook is effectively the internet for many Filipinos. News outlets, government agencies, and small businesses all depend on Meta’s platforms to reach audiences. The idea that Meta can now scrape and use their content without explicit, individual consent raises serious data privacy and competition questions.
The National Privacy Commission has been active on AI and data protection issues. The Data Privacy Act of 2012 gives Filipinos rights over their personal data, but it wasn’t written with AI training in mind. Meta’s commercial terms test the boundaries of what “consent” means in a world where platforms can unilaterally change the rules and present a binary choice.
For Filipino content creators and small publishers, the message is clear: the content you produce has value, and platforms will extract it whether you agree or not. The question is whether regulators will step in to ensure that consent is meaningful, not just a checkbox buried in a terms update.
What Can Publishers Do?
If you’re a publisher or business owner with a Meta presence, you have a few options — none of them great. You can accept the terms and hope Meta honors its stated commitment to respect robots.txt. You can reduce your dependence on Meta’s platforms by building direct audiences through email newsletters and your own website. Or you can join the legal fight to challenge the terms.
Some publishers are already updating their robots.txt files to block Meta’s AI crawlers explicitly. But as the legal experts point out, that may not be enough if Meta’s commercial terms are interpreted as overriding robots.txt. The technical solution may not hold up against the contractual one.
The most productive path is probably collective action. Industry groups representing publishers can push back on the terms, lobby regulators, and negotiate collectively with Meta. Individual publishers, especially small ones, don’t have the leverage to fight this alone.
The Bottom Line
Meta’s new commercial terms are a quiet but significant shift in who controls the web’s content. By framing the right to scrape as a condition of using its business tools, Meta has created a system where publishers’ work can be extracted for AI training without individual negotiation or compensation.
Meta says the change is about product recommendations, not news scraping. But the language says otherwise. And in a world where AI companies are paying hundreds of millions of dollars for content licensing deals, the idea that Meta can get the same content for free through a terms-of-service update should concern anyone who cares about a free and fair press.
The web’s content was never meant to be a free extraction zone for AI training. Whether it becomes one depends on whether publishers, regulators, and users push back — or whether they accept the terms and hope for the best.