Meta drops AI-training clause from business terms after news publishers object
Representative image: lawrence_monk / Pixabay
Meta has removed wording from its new commercial terms, due October 30, 2026, that let it use content from linked business websites to train AI, after publisher backlash.
What Happened
Meta has removed language from its updated commercial terms that would have let it use content from websites linked to a business's ads or products to train its artificial intelligence systems. The terms, which take effect on October 30, 2026, had said Meta's automated tools, such as crawlers and bots, could collect material from web pages and domains connected to a business using its products, and that Meta could use it to develop, improve, research and train its products. After news publishers and legal experts raised alarms in mid-October, Meta revised the text so the word "train" no longer appears and the crawling permission is tied to its advertising and business products. A Meta spokesperson said the edit does not change how its crawlers work and that site owners can still use robots.txt to control access.
Key Facts
- Meta's updated commercial terms take effect on October 30, 2026.
- The original wording covered automated tools such as scrapers, bots and crawlers running on pages and domains linked to a business's products or services.
- It said Meta could use what it gathered to develop, improve, research and train its products; no opt-out was offered.
- Legal experts told the journalism research site Nieman Lab the language could be read as permission to train AI on publishers' content, and could be argued to override a site's robots.txt instructions.
- Meta has now removed the word "train" and limited the crawling language to its advertising and business products.
- Meta says the change does not alter how its crawlers operate and that robots.txt can still be used to block them.
- Publishers and lawyers remain sceptical about the revised fine print.
Why It Matters
The fight over who may use web content to build AI models is one of the biggest unsettled questions in technology. AI companies need huge amounts of text to train their systems, and news publishers, whose reporting is costly to produce, want to be asked and paid before their work is used. Many publishers rely on robots.txt, a simple file that tells automated crawlers which parts of a site they may visit, to keep AI crawlers out.
The original clause worried publishers because almost every news site has a business relationship with Meta: they run ads, promote stories on Facebook and Instagram, or use Meta's business tools. If accepting Meta's commercial terms counted as consent to AI training, a publisher could, in effect, give up control over its archive simply by advertising its articles. Some experts warned Meta could even argue that a signed contract outweighed a robots.txt block.
By dropping the word "train", Meta has narrowed the clause on paper. But the permission to crawl linked domains remains, and critics point out that "advertising and business products" is a broad category at a company that increasingly builds AI into everything, from ad targeting to customer-service chat. That is why publishers say they will keep watching how the crawlers actually behave.
| What changed | Before | After |
|---|---|---|
| Use of crawled content | Develop, improve, research and train products | Tied to advertising and business products |
| Word "train" | Present | Removed |
| Permission to crawl linked domains | Yes | Still present |
| Effective date | October 30, 2026 | October 30, 2026 |
Impact
Short-term: Businesses and publishers that use Meta's advertising tools will accept the revised terms from October 30. Those who want to keep Meta's crawlers out can still set their robots.txt files to block them, according to the company.
Long-term: The episode shows that the fine print of everyday business agreements is becoming a new front in the battle over AI training data. Expect publishers, including in India, to read platform terms more closely and to push for explicit opt-outs or payment deals.
Who is affected: News publishers and website owners who advertise on Facebook and Instagram; small businesses that link their sites to Meta's business tools; and anyone following how AI companies collect training data.
Key Takeaway
Meta has removed explicit AI-training language from its new business terms, but it keeps the right to crawl linked websites, so publishers remain wary.
Questions and Answers
What did Meta change in its commercial terms?
Meta removed wording that said content crawled from websites linked to a business could be used to train its products. The revised text drops the word "train" and ties crawling to Meta's advertising and business products.
When do Meta's new commercial terms take effect?
The updated terms take effect on October 30, 2026.
Can website owners still block Meta's crawlers?
Yes. A Meta spokesperson said site owners can continue to use robots.txt to set their preferences, and that the edit does not change how Meta's crawlers work.
Why were news publishers worried about Meta's terms?
Legal experts said the original language could be read as permission for Meta to train AI on any content from sites linked to its business tools, and might even be argued to override robots.txt blocks.
Disclaimer: Prepared by the Peepals newsroom from publicly available sources with AI assistance. Information is accurate to the best of our knowledge at the time of publication and may change. Images may be representative. Not professional advice. Report an error via our contact page.
Sourced and fact-checked by the Peepals Global Editorial Team
Reported, fact-checked and published by the Peepals Global Editorial Team.
