Meta's New Terms Let It Crawl Business Websites for AI

Jeff's Daily Pick. Run a Facebook Page for your business? From October 30, Meta's contract says that also lets it crawl your website and train AI on it. There is no opt out.

In the United States and everywhere else Meta does business, a new version of its Commercial Terms takes effect on October 30, 2026. Nieman Lab reported on October 15 that the update gives Meta permission to send automated software to the websites of any business that uses its products, and to use what it collects to train those products. Three legal experts told the publication the clause could cover news sites, and virtually every newsroom has at least one Meta business account.

What the clause says

The preview of the terms, quoted by Nieman Lab, Ars Technica and Gizmodo, reads: "You authorize 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, analyze, collect, retrieve, or crawl the domains, web pages, or URLs (and any content available through them) that are associated with your products or services." A further sentence, quoted by The Register, lets Meta use that content to "provide, develop, improve, research, train, and maintain the integrity of" its products.

A business account is broad. Gizmodo lists accounts that advertise, sell, list products, manage a Page or business Group, or use the Meta Business Suite. The terms offer no opt out: a business that disagrees has to stop using Meta's commercial services. The phrase "test credentials" worried the lawyers most, because it could be read as permission to pass a login or a paywall.

Meta says its crawler behavior is not changing. The News/Media Alliance, which represents US publishers, called the policy "an appalling, bad-faith policy, and a clear abuse of the company's market power" in a statement on October 16. Whether a court would enforce such a clause against a copyright claim is untested.

The bill behind the crawler

The same week, analysts disagreed about the price of Meta's AI push. On October 14, Goldman Sachs cut its price target to $762 from $860, citing pressure on 2027 operating expenses, while raising its revenue estimates. UBS raised its target to $884 from $715, pointing to Business Agents and the Muse assistant as future sources of revenue in WhatsApp, Messenger and search advertising. Guggenheim moved to $850 from $700 and lifted its forecast for 2027 capital spending to $225 billion from $203 billion. Mark Zuckerberg has said Meta was ready to commit up to $200 billion to AI before knowing whether its new models would work. Meta reports third quarter results on October 21.

Why the two stories meet

Training data has become a cost line. Publishers have been signing licensing deals with AI companies, and a contract that obtains the same content as a condition of advertising would remove that cost for Meta. That is why publishers read the clause as a pricing decision as much as a legal one. The next dates are fixed: the terms switch on October 30, and investors hear the spending plan on October 21.