Meta may use sweeping new commercial TOS to avoid millions in AI licensing fees

Ars Technica reports Meta will update commercial terms on October 30 to require business accounts to allow Meta to train AI on linked website content, not only social posts. There is no opt-out short of stopping commercial use of Meta products. Nieman Lab notes virtually every news organization has at least one Meta business account. Attorney Matt Topic told Nieman Lab the buried provision would give Meta access to material that would likely cost many millions in a licensing market.

Ars compared old and new language and argues Meta’s “streamlining” story underplays fresh breadth, including “test credentials” for retrieving information from provided URLs. Meta tells outlets there is no change to how it uses web crawlers and that robots.txt still works. Prior crawling language in ad-related terms was more narrowly tied to ads.

Same story as The Verge’s Nieman Lab piece, with sharper teeth on the licensing math. If courts eventually bless broad AI training as fair use for some content, terms like these become the backup permission slip for everything else — especially news.

Useful AI needs training data. Extracting it by tying the social distribution pipe to an all-products crawl right is a choice, not a law of nature. Publishers already living with Meta’s traffic and Meta One link fees will feel this as another turn of the screw.

I’m watching Nieman Lab’s outreach to publishers, any state AG or EU interest in tying/adhesion theories, and whether Meta narrows the public FAQ before October 30.

Meta’s blog framed the update as reducing overlap between agreements. A Seller Agreement previously mentioned test credentials for e-commerce listing optimization — a narrower purpose than the new commercial terms’ training language.