Meta has pushed back the date its revised Commercial Terms and Self-Serve Ad Terms take effect, from 30 October to 20 November. A spokesperson told Reuters and Fast Company the company wants more time to explain what is changing and what is not. Meta also plans to describe how its web crawlers are used and to make clear that publishers can set their preferences through robots.txt.
“News publishers will be forced into providing their valuable journalism to support the development of products that may be used to compete with them for readers and advertisers, under duress and without compensation”
News Media Alliance, in an open letter to Meta · engadget.com ↗
The dispute centres on a clause that lets Meta use scrapers and other automated tools to access, index, cache, analyze, collect or crawl domains and pages linked to business accounts. The first wording also covered using that material to develop and train Meta products. After publishers complained, the word "train" was dropped and the purpose narrowed to advertising and business products. Most large publishers hold Facebook or Instagram business accounts, which is how their sites could come under the terms.
Whether the fix is complete is unclear. PPC Land found a UK-English preview of the 30 October terms that still held the original training sentence, and ContentGrip's own check of Meta's preview also turned up wording about training. Which locale or account sees which text has not been established. Meta maintains that the update concerns only its advertising and business products, that its crawler practice is unchanged and that it gives no new right to override robots.txt.
Featured graphic on the conflicting versions of Meta's AI-training terms ahead of the original 30 October date
Image: ContentGrip
Critics are unconvinced. The News/Media Alliance sent an open letter asking Meta to withdraw the terms before 30 October. Legal experts told Nieman Lab the text still gives Meta wide scraping permission over news sites, with Cornell's James Grimmelmann calling the business-product permission broad and attorney Matt Topic questioning whether it could stretch to AI tools. Robots.txt is voluntary, and Reuters has reported that AI crawlers often ignore it. Complaints have also gone to antitrust authorities.
About Meta
Meta Platforms, doing business as Meta, is an American technology company headquartered in Menlo Park, California. It owns and operates Facebook, Instagram, WhatsApp, Messenger and Threads, and runs an advertising network. It began as Facebook and rebranded as Meta in 2021 to reflect a shift toward the metaverse.
Meta has been pulling back from paying news publishers for some time. In March 2024 it said it would wind down Facebook News in the US and Australia and stop negotiating new commercial deals with publishers. On 5 December 2025 it announced commercial AI data agreements with publishers including USA Today, People Inc, CNN, Fox News, The Daily Caller, Washington Examiner and Le Monde, with terms undisclosed.
7
Models indexed
Models on AIPopDB
Meta stopped paying publishers for news in 2022. The terms now take effect on 20 November.
Sources · 3