Meta has removed disputed wording from its commercial terms after experts interpreted it as potentially allowing website content to be used to train its AI. The company says the change does not alter how its web crawlers work, leaving a meaningful clarification but not a new promise about how crawling operates.
Meta AI Watch analysis
What happened
The wording, which referred to using content from domains, web pages or URLs to develop and train Meta products, has been removed from the terms, The Verge reports. Meta spokesperson Daniel Roberts said the terms applied to ads and business products, and that the revision clarifies this. He also said the change does not alter Meta’s use of web crawlers and that site owners can use robots.txt to signal their preferences.
Experts, including Cornell law professor James Grimmelmann, had interpreted the earlier wording as potentially giving Meta rights to scrape site content for AI training and overriding robots.txt instructions. Meta disputes that interpretation. The The Verge report sets out both positions.
Why it matters
Terms pages are not thrilling bedtime reading, but this one sits at the intersection of AI training data, web crawling and the choices publishers can signal about their content. Meta’s change removes wording that prompted concern; its spokesperson says it did not change the company’s crawler practices or give it new rights.
That distinction matters. A revised contract can clarify what a company says its terms mean without settling every question about how its systems behave in practice.
Our read
Removing the contested language is a useful clarification, not proof that the underlying dispute has vanished. Meta’s account is clear: the terms concerned ads and business products, and robots.txt remains the way for site owners to signal preferences. The important test is whether the revised terms and actual crawling behaviour continue to match that explanation.
What to watch
- Whether Meta publishes further detail about the scope of its crawling terms.
- Whether publishers or legal experts accept the revised wording as addressing their concerns.
- Whether Meta’s stated treatment of robots.txt changes or is challenged.
Discussion spark: Should publishers treat Meta’s revised wording as a meaningful safeguard, or does the real test have to be what its crawlers do with robots.txt?
Sources and evidence
- Source update (17 October 2026, 15:04 UTC)
not affiliated with, endorsed by, or operated by Meta