Meta has removed terms language that critics said could let it scrape website content to train AI models.
On 17 October, The Verge reported that Meta deleted wording from its commercial terms after backlash over language that experts said appeared to permit scraping content from websites, including news sites, for AI training. The removed text had said Meta may use content from domains, web pages or URLs to provide, develop, improve, research, train and maintain Meta products and services.
TechCrunch reported that legal experts interviewed by Nieman Lab had described the earlier wording as a potentially significant provision. Meta, however, said the revised terms do not reflect a change in how it crawls the web, and that the company has now clarified the language in the terms themselves.
What changed in Meta's terms
According to The Verge, the wording that triggered concern no longer appears on Meta's commercial terms page. That matters because terms pages set the rules for how a company says it can use data or content connected to its products.
In this case, the disputed clause referred to content from linked domains, web pages and URLs. Critics read that as broad enough to cover material on publishers' sites and other websites tied to Meta business accounts. In simpler terms, the concern was that a business using Meta tools might be seen as granting permission over content on its website, with possible implications for AI training.
Why experts and publishers objected
The Verge said experts interpreted the clause as an attempt to give Meta the right to use scraped content to train AI products without payment, and potentially to override blocking instructions in a site's robots.txt file. Robots.txt is a simple file website owners use to signal whether automated crawlers should access certain pages.
TechCrunch noted that legal experts interviewed by Nieman Lab viewed the language as a "remarkable provision" because it appeared broad enough to let Meta scrape the website of any publisher with a Meta business account and use that material to train AI models.
Those interpretations do not mean Meta had already changed its crawler behaviour. But they do show how quickly contract language can become a governance issue when companies are using or developing AI systems that depend on large volumes of online material.
What Meta says the language did and did not do
Meta's position, as quoted by both TechCrunch and The Verge, is that there was "no change" to how it uses web crawlers. The company said the consolidated crawling terms applied only to its ads and business products, not to a new effort to scrape news sites.
The Verge also reported Meta's further statement that content owners can still use robots.txt to signal their preferences, and that the updates do not give Meta new rights that supersede those preferences. That is an important limitation in the company's account: Meta is framing the edit as a clarification and removal of confusing language, not as a reversal of a newly claimed technical capability.
The gap between the outside interpretation and Meta's explanation remains central to the story. Based on these reports, what changed for certain is the text of the terms. What is not established by the sources is any proven change in Meta's actual crawling operations before or after the edit.
Why the wording matters more than the crawler debate
The strongest conclusion supported by these reports is narrow but important: Meta changed the text of its terms after criticism, while insisting its actual web-crawling practices did not change. That means the clearest confirmed development is contractual and communicative, not technical.
What matters most is how easily ambiguous language about domains, URLs and content use can be read as a claim to broader AI training rights. For publishers, website owners and businesses using platform tools, the practical lesson is to watch terms updates closely, especially where crawling and AI training are mentioned together.
The sources do not provide direct evidence that organisations changed purchasing or platform-use decisions in response. What they do show is that clearer terms language is the main immediate takeaway, because that is the part that changed and the part Meta publicly addressed. Whether similar disputes continue will depend on how large platforms describe content-use boundaries in future updates.