Meta deleted one very spicy word from its upcoming business terms on Oct. 16. The company made the edit just one day after Nieman Lab reported that legal experts believed the fine print could let Meta scrape news sites to train its AI systems, according to The Verge. Did anyone at Meta bother reading the thing before hitting publish? The word was “train,” and while it is gone, the permission to crawl websites tied to business accounts is still very much there.
The Clause That Launched A Thousand Headaches
The original sentence said Meta “may use any content from such domains, web pages, or URLs to provide, develop, improve, research, train, and maintain the integrity of” its products and services. That is a lot of verbs for one sentence.
Legal experts who reviewed the language for Nieman Lab, including Cornell Law School professor James Grimmelmann, warned it could let Meta train AI on scraped content for free and argue that its contract overrides the robots.txt instructions websites use to block automated crawlers, according to The Verge. My read is that Meta wanted free AI training data, and the wording was doing a lot of heavy lifting to make that happen.
Why This Mattered More Than Your Average Terms Update
Virtually every news organization maintains a Meta business account for pages, ads or marketing tools. So this was not some obscure corner of the internet affecting 12 people and a guy named Gary. The new commercial terms apply to business accounts, not just personal social media posts, and grant Meta rights to crawl the websites linked to those accounts. That scope reaches way past your cousin posting photos of his smoker. Businesses also cannot reject the new terms and keep using Meta’s services, according to Ars Technica.
Matt Topic, a media and intellectual property attorney at Loevy & Loevy, said the provision stood out for where it was placed. He called it “a remarkable provision to bury in terms of service” and pointed out the material would likely cost many millions in a licensing market. Burying something like that in the terms is like hiding a motorcycle in your carry-on and hoping nobody notices at security.
The Paywall Problem Nobody Asked For
The terms also let Meta create something called “test credentials” to access those websites. Grimmelmann said that could let Meta register for sites to slip past registration walls, and Meta declined to tell Nieman Lab whether the credentials could get around paywalls. Ars Technica found the phrase was nowhere in the older commercial terms. And unlike the word “train,” “test credentials” survived the edit. Cool. Totally normal. Nothing to see here.
Experts warned Meta could train AI on publisher content and then deploy that AI to compete directly with the same news organizations supplying the material, according to Ars Technica. That concern lands right as Meta has started charging some users to post links through its Meta One subscription, as Nieman Lab pointed out. So Meta charges to post links to the same content it wants to crawl for free. Everything is fine. The vibes are immaculate.
Meta Says Nothing To See Here, Folks
Meta disputed that anything substantive had changed in how it collects data. Andy Stone, the company’s vice president of communications, said: “There is no change to how Meta uses web crawlers; any suggestion to the contrary is false.” Daniel Roberts, a Meta spokesperson, offered a longer explanation. He said the crawling terms “only applied to our ads and business products” and that content owners can still use robots.txt to signal their preferences.
Roberts also said the updates “don’t give us new rights that supersede those preferences.” Sure. And I am the Queen of Sheba. My opinion is simple and I am sticking to it. Meta is a company that scrapes first and apologizes later. This whole episode did nothing to change my mind.
The lawyers were not exactly soothed either. Grimmelmann told Nieman Lab the revised wording still leaves the “associated with your products or services” language untouched, which he said would let Meta scrape entire sites.
The Questions Meta Would Rather Skip
Meta’s statements address how the company says it currently uses web crawlers. They do not explain why the broader language appeared in the first place. The promise to respect robots.txt also lives in a spokesperson’s statement, not in the contract itself. That is a pretty big difference when lawyers get involved.
It also remains unclear how this episode affects licensing discussions between Meta and news organizations over AI training data. Publishers are left relying on Meta’s public statements and their own robots.txt settings. That is like relying on a fox to tell you which chickens are safe.
What Publishers Do Now
The new terms were set to take effect Oct. 30. Meta pushed that date to Nov. 20 after publishers, advertisers and press groups complained, including to EU and UK antitrust regulators, Reuters reported. The delay changed the calendar, not the text, and Meta has not said whether more edits are coming. The whole situation has the energy of a kid caught with a hand in the cookie jar who insists the jar was already open and the cookies were stale anyway.
A Few Lingering Thoughts
The funniest part is that this all came to light because a reader tipped off Nieman Lab. This is the fine print equivalent of finding out your landlord slipped a clause into your lease saying he can use your living room for band practice. Legal experts flagged the wording. Publishers raised alarms. Meta deleted one word and hit snooze for three weeks.
The only thing publishers can do before Nov. 20 is keep an eye on those terms pages and hope the next round of fine print does not include a clause about getting a free pony. I have said it before, and I will say it again. Read the terms. All of them. Even the boring parts. Especially the boring parts.
More Gear & Tech news from Total Apex Entertainment.
Don't stop here
Every Score. Every Story. One Home.
Breaking sports, entertainment and news from across the Total Apex network, updated all day.
Get the Full Story →
