Meta Drops Scraping Clause From Business Terms of Use

A clause in the company's commercial terms of use that let it pull content from any business partner's website to train its artificial intelligence systems has been deleted, according to a report from TechCrunch, which cites reporting from The Verge on the removal.

What did the original language say?

The updated terms had declared that the social media giant could "use any 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, TechCrunch reported. The wording applied broadly to any business with an account on the platform, not just advertisers running campaigns.

Why did publishers and lawyers object?

Legal experts interviewed by Nieman Lab called the provision a "remarkable" grant of authority, warning it could be read to give the company permission to scrape the site of any publisher with a business account and feed that material into its AI models, according to TechCrunch's account of the Nieman Lab analysis. The concern carried extra weight because litigation over AI training data remains active in several courts even after the company won a high-profile copyright case last year, TechCrunch noted.

What does the company say actually changed?

A spokesperson told TechCrunch the terms never signaled a shift in crawling practice. After the language was pulled, the company issued a statement reported by TechCrunch: "We've heard you and want to clarify: the crawling terms we consolidated only applied to our ads and business products, and we've clarified that in the terms themselves. It does not change how Meta uses web crawlers."

The statement draws a distinction between data tied to advertising and business tools -- information publishers hand over when they run campaigns or manage a page -- and the open-web crawling the company uses to gather training data more broadly. The company's position is that the clause was a consolidation of existing ads-and-business terms, not a new crawling authorization, and that removing it was a clarification rather than a policy reversal.

Timeline of the dispute

• The commercial terms of use were updated to include the training-data language, drawing notice from Nieman Lab's legal sources.
• TechCrunch published the concerns on October 17, 2026, citing the Nieman Lab analysis.
• The Verge reported the clause had been removed following the scrutiny, according to TechCrunch's follow-up item.
• The company issued its clarifying statement the same day the removal was reported.

What to watch next

• Whether the company publishes further detail distinguishing ads-and-business data from general web-crawling practice.
• Whether other AI-training lawsuits against the company, still pending despite last year's copyright win, cite the deleted clause as evidence of intent.
• Whether competing platforms revise their own business terms after the scrutiny this case drew.
• Whether publishers request explicit opt-out language rather than relying on verbal assurances.

The episode lands alongside broader workplace shifts tied to AI adoption, a theme explored in The Economist's Bartleby column on aging workers and AI, and underscores how terms-of-service language has become a flashpoint in the AI-training debate even when a company insists no practical change occurred.