Meta’s new terms may allow it to train AI on news sites for free
Oct 16, 2026, 7:02 AM · The Verge
Meta’s commercial terms update, effective October 30, lets it crawl sites “associated” with business accounts for AI training — publishers hear a power grab.
Why it matters
The Verge, citing Nieman Lab, reports Meta’s new commercial terms — effective October 30 — authorize Meta to use scrapers and crawlers on domains and URLs associated with a business’s products or services, and to use that content to develop, improve, research, and train Meta AI products. The terms apply to business accounts promoting across Facebook, Instagram, WhatsApp, and Messenger. Businesses that won’t agree may lose commercial access to Meta products.
Meta spokesperson Daniel Roberts says the company is consolidating commercial and self-serve ad terms and that content owners can still use robots.txt. Legal experts tell Nieman Lab the terms could let Meta argue commercial terms supersede blocking instructions. Cornell’s James Grimmelmann says for news publishers this looks like permission to train on site content and possibly deploy it in competition. Andy Stone of Meta says the update is about accurate product info for ads and recommendations, not scraping news sites.
From the desk
Believe the text of the contract over the soothing quote. “Train” is in the terms. “Associated” websites are in scope. Opting out means walking away from Meta’s business surfaces — which for many publishers is not a real choice.
Meta already struck licensing deals with outlets including CNN, Fox News, USA Today, and People Inc. These terms look like a wider net underneath those deals. Pair that with Meta One charging for links in posts and reels, and Nieman Lab’s framing stings: pay to post the URL, then grant a crawl of what it points to.
We’re not anti-crawler when the use is clear and consented. This is consent by adhesion. I’m watching whether publishers pull business accounts before October 30, and whether robots.txt remains respected in practice once the new terms are live.
Context
Ars Technica’s parallel reporting flags new “test credentials” language that experts worry could help bypass paywalls, and notes Meta’s claim that crawling language was previously scattered across policies.
Who feels it
- News publishers
- Pressure to accept AI-training crawl rights or exit Meta business tools.
- Small businesses
- Broader site crawl permissions tied to ordinary ad and shop use.
- Licensing negotiators
- Harder to price deals if TOS already grants wide training access.
What to watch
- Publisher exits or collective pushback before the October 30 effective date
- Whether Meta’s robots.txt assurances hold after the terms change
- Any enforcement fight over “associated” domains beyond product pages
Companies: Meta