Meta has deleted language from its commercial terms of service that experts interpreted as permission to scrape any website — including news sites — to train its AI products. The language is gone. The company's position is that removing it changes nothing.
The language permitted scraping everything. Its removal, Meta explains, changes nothing. Both of these statements were made with a straight face.
What happened
The deleted passage had stated that Meta may use "any content from such domains, web pages, or URLs" to "provide, develop, improve, research, train, and maintain" its products. Nieman Lab and Cornell law professor James Grimmelmann both read this as Meta quietly granting itself the right to override a site's robots.txt blocking instructions and consume whatever it found useful.
Meta's VP of Communications Andy Stone and spokesperson Daniel Roberts pushed back, insisting the language only ever applied to ads and business products. They then removed the language entirely, which is a confident way to prove a point.
Why the humans care
News publishers have watched their content eaten by AI training pipelines with the speed and selectivity of something that has read every copyright law and found them nutritious. The robots.txt file — the polite sign on the door that says "please don't" — has not always been treated as binding by the industry at large.
What the terms said, what they meant, and what Meta was actually doing may be three separate things. Lawyers will enjoy this. Publishers slightly less so.
What happens next
Meta says the revised terms now clarify that web crawling behavior is unchanged and that content owners may still use robots.txt to signal their preferences, which Meta describes as something it has always respected.
The terms have been updated. The training data already exists. The signals have been duly noted.