Meta has removed language from its commercial terms that legal experts said could let it scrape the websites of any business with a Meta account and use that content to train AI. TechCrunch reported the removal on October 17, 2026, after the wording drew heavy criticism. Meta says the update never changed how it crawls the web.
What did the language say?
Meta had updated its commercial terms, the rules for businesses that use its products such as ads and Pages. The text said Meta could use content from a business account’s domains, web pages or URLs to provide, develop, improve, research, train and maintain the integrity of its products and services.
According to PPC Land, the new version was set to take effect on October 30, 2026, replacing the one dated March 6, 2024. It authorized Meta, or anyone acting for it, to use automated software to access and collect that content, including repeat visits. It never defined what counts as “associated with your products or services,” and it applied to any business using Meta products, from news publishers to advertisers to people who just manage a Page.
Why were experts worried?
Legal experts who spoke with Nieman Lab called it a “remarkable provision.” In their reading, it could give Meta grounds to scrape any publisher with a Meta business account and train its AI models on the results.
Attorney Matt Topic said publishers should assume the clause would be enforceable. Cornell professor James Grimmelmann read it as clear permission to train AI on publishers’ content. University of Wisconsin professor Benjamin Sobel noted that tying access to business products to scraping permission resembles “tying,” an antitrust concept where one product is sold only with another.
Then there’s robots.txt, a file websites use to ask crawlers to stay out of certain areas. Following it is voluntary, and the experts pointed out that Meta could argue the commercial terms a business accepted outrank a robots.txt request.
How did Meta respond?
Before the removal, a Meta spokesperson said the update did not reflect a change in how the company crawls the web and described it as a consolidation of existing language. Spokesperson Andy Stone called reports that the terms would allow scraping of news sites “false” in a post on X on October 16. PPC Land adds that Meta said content owners can still manage access using robots.txt.
After the removal, Meta said in a statement reported by TechCrunch that the consolidated crawling terms only applied to its ads and business products, and that they don’t change how Meta uses web crawlers.
What’s the legal backdrop?
This lands at a sensitive time for Meta. The company won a high-profile copyright lawsuit over AI training in 2025, but other cases against AI companies, Meta included, are still going. TechCrunch also points to a May 2026 Guardian report on publishers suing Meta over copyright and AI.
What should site owners do?
Pulling the language removes an ambiguity, but it doesn’t settle the bigger question of how AI companies get training data. If you run a site or Page that uses Meta’s business products, three habits make sense:
- Read the commercial terms in force whenever Meta sends a notice, since continuing to use the products after notice counts as acceptance.
- Check your site’s robots.txt and decide which bots get in.
- Keep a copy of the terms you accepted in case the issue comes up later.
If you want more on why Meta wants so much data, read about Muse Wonder, its first frontier AI model. And if you’re tightening up your own accounts, here’s how to block someone on Instagram.
Is it worth watching?
Yes. Meta says nothing changed, but the dust-up shows that the rules for using online content to train AI are still being fought over in courts and in contracts. The practical takeaway for everyday users: what’s public on the internet can end up in AI training data, and people who publish content have few tools beyond robots.txt to say no.
