Meta rewrites commercial terms to scrape business websites for AI training
October 30 update grants no opt-out for publishers whose content would cost millions in a licensing market, Meta says it has no plans to ignore robots.txt
Images
Photo of Ashley Belanger
arstechnica.com
Meta will update its commercial terms of service on October 30 to require business accounts to let the company train AI on content from websites linked to their accounts — not just their posts. There is no opt-out. Businesses that decline are invited to stop using Meta services. Nieman Lab, which first reported the change, noted that virtually every news organization holds at least one Meta business account.
The updated terms authorize Meta to use automated scrapers, bots, spiders and crawlers to access, index, cache and collect material from domains associated with a business's products or services, and to periodically pull updates from them. The stated purpose covers developing, improving, researching and training Meta's products. The word 'AI' appears nowhere in the text; the permission is inferred from 'train' and the scraping language. Meta described the change in a blog post as 'streamlining' its commercial terms, and a spokesperson told Nieman Lab the company has no plans to use the terms as a legal basis to override robots.txt instructions that block crawling on news sites.
Three legal experts told Nieman Lab it is unclear whether a court would enforce the terms. James Grimmelmann, a professor of digital and information law at Cornell Law School, said Meta could train AI on any content from businesses' sites and possibly deploy it — in competition with the publishers themselves — on Meta platforms. Matt Topic, a media and intellectual property attorney at Loevy & Loevy, said the provision hands Meta material that would likely cost many millions of dollars in a licensing market.
That market exists because it is being contested elsewhere: news organizations are currently fighting OpenAI in court over whether AI training without licensing is fair use. Norwegian media companies have sued Meta over copyright, and a Swedish prosecutor is investigating a publishers' fraud complaint — while Swedish advertisers keep paying their invoices. The pattern is consistent. Take the content, pay nothing, and let a clause in unread terms do the work copyright law was supposed to do. Meta, for its part, recently began charging users to post links.