Meta drops AI-training language from ad terms as publishers push back

Meta has pulled the most contentious wording from the commercial terms it wants advertisers and business users to accept: the right to train AI on content crawled from their sites. The crawlers stay. The deadline does not move.

According to Unite.AI, an archived capture from 16 October shows the original clause, and a capture from the early hours of 17 October shows the revised text. The removed sentence began: "Meta may use any content from such domains, web pages, or URLs to provide, develop, improve, research, train". The verb "train" no longer appears in the provision.

The rewrite leaves Meta's access largely intact. The clause still lets Meta, or anyone acting for it, use "web scrapers, bots, robots, spiders, crawlers, and user agents" to "access, index, cache, analyze, collect, retrieve, or crawl" business users' domains and pages. The stated purpose is now to "provide, develop, improve, research, and maintain advertising and business products", and to maintain the integrity and security of Meta products.

Other edits trim the document. The licence section for shared content is shorter, and the sentence about helping users create and deliver ads is gone. The Self-Serve Ad Terms preview carries the same revised clause. Product Catalog Terms are now folded into the Commercial Terms, and Ad Creative Generative AI Terms into the Self-Serve Ad Terms. The revised terms still take effect on 30 October 2026.

Publishers reacted before the change surfaced. Danielle Coffey, president and CEO of the News/Media Alliance, called the terms "outrageous and probably illegal" and said they could not be implemented on 30 October. Meta spokesperson Andy Stone said the update "was not about scraping news sites". He said there was no change to how Meta uses web crawlers, and that content owners can use robots.txt flags. Both statements came on 16 October.

One caveat: Unite.AI infers the edit from archive captures. Meta has not said it changed the text.

Meta has removed the phrase that drew the fire without giving up the access behind it. The new wording still allows broad crawling and collection of business users' pages, and "develop" and "research" for advertising and business products remain in the purpose clause. Whether that covers AI work is a question the text no longer answers directly, which is a choice in itself.

Publishers hold the exposed position. Stone points them to robots.txt flags, which puts the burden of opting out on content owners. The News/Media Alliance has said the terms cannot be implemented on 30 October, and Meta has kept that date. Watch whether the Alliance and others treat the edit as enough, and whether Meta ever confirms the change in its own words.