Meta Removes the Word "Train" From Its Business Terms, but Critics Are Not Satisfied

Meta has removed the word "train" from its business terms. The old wording said the company could use content taken from news sites to train its products. The revised terms for business accounts take effect on October 30, according to The Verge.

Scraping means using automated software to copy material from websites. Training an AI product means feeding it large amounts of material so it learns patterns from that material.

The earlier wording said Meta "may use any content" from websites linked to a business account to "develop, research" and "train" its products. The language drew criticism once the terms came to light. The Verge reported on October 16 that the terms seemed to give Meta permission to scrape online publications for free, citing Nieman Lab's analysis. A follow-up on October 17 reported that Meta was updating the terms again to remove language suggesting it could scrape content, including from news sites, to train its AI.

Not everyone is satisfied. Cornell Law professor James Grimmelmann said the revised terms do nothing to limit the phrase "associated with your products or services." He argued that this phrase would still let Meta scrape entire sites. Media attorney Matt Topic said the revised terms still allow Meta to use scraped content to develop any "business product." In his view of the wording, "develop" is broad enough to include "train."

Meta spokesperson Daniel Roberts said content owners can use robots.txt to signal their preferences. He said the update "is not about scraping news sites."

Robots.txt is a short text file a website can publish to tell automated programs which parts of the site to stay away from. It works like a "keep out" sign on a door. A visitor has to choose to respect the sign.

Nieman Lab went further on how consent would work. It reported that business accounts that keep using Facebook, Instagram, or WhatsApp after October 30 would agree to Meta scraping their websites under the updated Commercial Terms.

In my view, the dispute comes down to exact wording rather than the headline change. Removing "train" changes what the document says on its face. Whether "develop," combined with "associated with your products or services," leads to the same result in practice is the question Grimmelmann and Topic are answering. They answer it differently from Meta's spokesperson. Anyone who manages a website should read the sections that control what happens, not rely on summaries.

The broader context here is the opt-out system. Robots.txt is voluntary, so a publisher can ask certain crawlers to stay away, but the crawler has to agree to follow the request. It also does not control, on its own, what a platform does with content it was already allowed to access under a contract. Meta's emphasis on robots.txt puts the job of stating preferences on the content owner. That is a different starting point from a rule that requires permission first.

The October 30 date gives publishers and business account holders a short window to read the terms, decide whether to keep using the services, and update their sites. If the terms take effect as planned, the final wording will matter most. The arguments over "develop" and "associated with" would then move from commentary into any enforcement or legal action that follows. For now, the concrete change is one removed word and a date on the calendar.