Meta is set to push through commercial terms that legal observers describe as a quiet land grab over the content of every business that uses its services.
The update takes effect October 30 and requires business accounts to permit Meta to train AI on content from domains linked to those accounts, not just the posts they publish on Facebook or Instagram. There is no opt-out. The terms suggest that any business unwilling to agree should stop using Meta’s products.
The clause that does the work
The language authorizes Meta or anyone acting on its behalf to use scrapers, bots, spiders, crawlers and user agents, and to create test credentials, to access, index, cache, analyze, collect, retrieve or crawl associated domains and their content. That content may then be used to develop, improve, research and train Meta’s products.
Notably, the word AI does not appear. The permission is inferred from a single use of the word train alongside the scraping language.
Why publishers are alarmed
News organizations are already litigating whether training on their work without a license is fair use. As Nieman Lab noted, the terms would seem to grant Meta sweeping permission regardless of how those cases resolve. Virtually every newsroom holds at least one business account.
Meta’s spokesperson told Nieman Lab the company has no plans to use the terms as a legal basis to ignore robots.txt instructions. The terms themselves make no such promise, and the licensing fees at stake for publishers run into the millions.
