Meta's new commercial terms may let it scrape publisher sites for AI training, Nieman Lab reports

Left · 1Nieman Lab reported that Meta's updated commercial terms appear to let it crawl and scrape websites associated with businesses that use its platforms, including online publishers. 1 Left · 1The terms take effect on October 30 and apply to business accounts that promote content across Facebook, Instagram, WhatsApp and Messenger. They say Meta may use content from those domains to "develop, improve, research" and "train" its products. 1 Left · 2The clause authorises Meta or anyone acting for it to use scrapers, bots and crawlers, and to create "test credentials", to access and pull updates from linked domains. Ars Technica notes the text never mentions AI; that meaning is inferred from the word "train" and the references to automated scraping. 1 2

Left · 2Businesses that do not accept the terms "may stop all access or use of our Products for business or commercial purposes". Ars Technica says the update does not allow businesses to opt out. 1 2 Left · 2Meta has struck AI licensing deals with CNN, Fox News, USA Today and People Inc., so the terms could give it access to far more content than those deals cover. Legal expert Matt Topic called the provision "remarkable" to bury in terms of service, because the material would likely cost many millions of dollars in a licensing market. 1 2 Left · 1Last month Meta launched a Meta One subscription that requires businesses and creators to pay to include links in posts and reels. Nieman Lab said Meta has begun charging users to post links to content it will be permitted to crawl for free. 1

Left · 1Three legal experts told Nieman Lab it is unclear whether a court would enforce the terms, but that they could be used to argue Meta may scrape news sites for AI training. Cornell Law professor James Grimmelmann warned Meta could train on any content from the sites and possibly deploy it in competition with publishers. Experts also likened the arrangement to the antitrust concept of "tying". 2 Left · 1Meta described the change as "streamlining", saying crawling language was previously scattered across different terms and is now consolidated. A spokesperson told Nieman Lab that content owners can still manage access through robots.txt, and told Ars that "there is no change to how Meta uses web crawlers; any suggestion to the contrary is false." 2 Left · 1Ars Technica compared the documents and found that "test credentials" appear in neither the current commercial terms nor the Product Catalog or Ad Creative Generative AI Terms being merged in. It says the wording seems to have been introduced this month in the Self-Service Ads Terms, and that the earlier crawling language covered only ad content. Meta did not immediately respond to Ars' request to clarify. 2

Left · 1News organisations are already contesting with OpenAI whether training without a licence is fair use, and publishers fear the terms could override robots.txt blocks if courts rule for AI developers. Stephen Beale, founder of The Steampunk Explorer, told Nieman Lab he is considering leaving Facebook over the update. 2