Meta removes disputed clause on website data and AI training

A broad clause, then a narrower explanation

The removed language said Meta could use content from domains, web pages and URLs associated with a business account to provide, develop, improve, research and train its products, and to support their integrity.

Lawyers interviewed by Nieman Lab called the provision unusually broad. They said it could have given Meta the right to collect material from any publisher’s website if that publisher had a Meta business account, and use it to train AI models.

Meta’s representative had said the updated terms did not reflect a change in how the company scans the internet. The Verge later reported that Meta had removed the disputed provision.

In its statement, Meta said it had heard the feedback and clarified that the combined scanning terms applied only to its advertising and business products. The company said the clarification did not change how it uses web crawlers.

What the deletion settles — and what it doesn’t

The wording matters because Meta won a high-profile copyright lawsuit last year, while other cases against AI companies, including Meta, are still ongoing. But the company’s explanation is about the scope of its terms, not a change to its crawling practices.

I think that distinction is the point: removing a clause can reduce ambiguity without answering what material Meta’s crawlers collect or how it is used. The announcement says the practice is unchanged, but offers no further detail about that practice. For publishers, the dispute therefore shifts from what the contract appeared to authorize to what Meta actually does.