Meta pauses its AI scraping terms indefinitely after courts in two countries and a UK investigation

Meta has paused proposed commercial terms that would have let it collect content from the websites of businesses and publishers that use its platforms and use it in its own products, including AI features. The pause, reported on November 5, 2026, came after two foreign courts, a UK competition investigation and a letter from two US senators. It is a retreat in timing, not yet in substance.

What Meta proposed

Nieman Lab first reported the terms on October 15, 2026. They covered Meta's Commercial Terms, Product Catalog Terms, Self-Serve Ad Terms and Ad Creative Generative AI Terms, and legal experts told Nieman Lab they would allow scraping of news sites. The CMA described the same risk in its own words: the terms could let Meta collect content from business and publisher websites and use it for its own products and services, including AI features.

Meta moved quickly under criticism. On October 16 it dropped the word "train" and limited use of scraped content to advertising and business products, per Nieman Lab. On October 23 the original October 30 effective date moved to November 20. On October 28 Andy Stone, Meta's vice president of communications, wrote that the company did "not plan to make policy changes without a full consultation with industry groups and leaders."

The pause

On November 5, Meta removed the preview page and updated its Business Help Center. The notice reads: "These updates are now paused and will not take effect on November 20, 2026." A spokesperson told Nieman Lab: "We’re pausing the changes as we continue to engage with stakeholders and take our time to get this right."

Existing terms stay in force. Nieman Lab does not report a new date, and the pause is described as indefinite. The CMA's page describes Meta's October 28 step as a pause pending consultation, so the exact sequence of pauses differs between sources. What is consistent is that nothing takes effect on November 20.

Three kinds of pressure

First, courts. Nieman Lab reports that courts in Norway and Brazil issued preliminary injunctions blocking the terms following lawsuits by media organizations. Preliminary injunctions are early rulings, not final findings, but they mean the terms were already unenforceable in two markets.

Second, the CMA. Its press release of October 30, 2026 says it is investigating under the Competition Act 1998 whether Meta holds a dominant position in a relevant UK market and, if so, whether the terms abuse it. The CMA suspects Meta may be imposing exploitative terms on publishers and advertisers, and that requiring them to accept the terms could amount to unfair trading conditions. It sets no deadline and says no assumption should be made that the law has been broken. Sarah Cardell, the CMA's chief executive, said: "We will be working to reach a swift and satisfactory resolution to this case."

Third, Congress. Senators Adam Schiff, a Democrat from California, and Marsha Blackburn, a Republican from Tennessee, wrote to Mark Zuckerberg asking him to withdraw the terms, per Nieman Lab. A two-party letter on an AI data question is rare enough to note.

Why it matters beyond Meta

The central question is whether a platform can condition ordinary business services on permission to reuse a customer's content for its own AI products. Training-data fights have so far been about crawlers that ignore signals. This one is about contract terms that would make consent a condition of buying ads or running a catalog.

That is why the UK investigation is the one to watch. It tests the idea that bundling can be an abuse of dominance, and the CMA's separate work under its digital markets regime is due before its Board in spring 2027, though Meta is not currently designated. Our tracker of AI training opt-out signals shows how few of these rights are settled elsewhere, and the shared robots.txt token piece explains the voluntary route that this contract approach sidesteps.

What the CMA is and is not doing

The CMA says it is separately seeking information on Meta One and on concerns about fees for external links on Facebook and Instagram. Those fall outside this investigation, and no decision on formal action has been made on them. The agency also noted that Meta is not currently designated under the digital markets regime, so this case runs under general competition law rather than the new conduct rules. That choice has consequences: an abuse of dominance case needs a market definition and evidence of dominance, which can take longer than a conduct requirement.

Meta's original plan was to change its terms on October 30. By the time the CMA announced its case that same day, the company had already moved the start date, which shows how fast the commercial pressure was working compared with the legal process.

What to watch

• Whether Meta publishes revised wording and, if so, whether it restores any AI-product use of scraped content.
• Whether the CMA sets a timetable or accepts commitments.
• Whether the Norwegian and Brazilian injunctions become final rulings.
• Whether other platforms test similar terms while Meta is paused.

Our take

The pause is the right outcome and the wrong place to stop. A publisher should not have to choose between running ads and licensing its archive to an AI product, and consultation after the fact is a weaker remedy than a clear rule. We would watch the CMA most closely: a finding on exploitative terms would set a UK reference point for every platform drafting similar language.