Meta drops the word "train" from its commercial TOS after AI backlash
Ars Technica Ashley Belanger ● Covered by 9 sources
Meta quietly removed “train” from its business terms after backlash over AI scraping. It doesn’t clear up the policy; it may just make the whole thing harder to read.
Based on reporting by Ars Technica, Ashley Belanger — read the original for the full story.
Summary, retelling and take written by AI under human oversight; images are AI-generated illustrations. How we work · Report an error
Meta has changed a key word in its commercial terms of service after a wave of criticism over what the policy might allow. The word “train” is gone from the “streamlined” terms that are set to take effect on October 30.
The trouble started when legal experts said last week that Meta’s new language looked broad enough to let the company scrape websites linked from business pages over and over again. That raised the possibility that Meta could use the policy to dodge licensing costs tied to news content, with business users effectively handing over AI training material just by keeping a page active.
Meta’s response so far has been limited. A spokesperson confirmed that users were not alerted to the language change after the backlash, which hardly helps the confusion around a policy that is already drawing close scrutiny before rollout.
And the company’s statement to Ars only went so far as to acknowledge Nieman Lab’s first report. It did not address the follow-up that repeated experts’ concerns about the policy’s reach. When a term of service starts looking like a legal puzzle box, removing one word is not much of a fix.
My take — AI-written commentary, not fact-checked reporting
Meta keeps finding new ways to make “commercial terms” sound like a hostage note. Dropping one word after the blow-up is classic big-platform behavior: change the wording, keep the power, hope everyone gets tired. The real issue is simple — if a policy needs a decoding team, it’s probably doing too much for the company and too little for everyone else.
Read more about this at: Ars Technica