started · updated
Meta removes AI training language from business terms after backlash
Meta has revised its commercial terms following significant criticism from legal experts and news publishers. The original proposed language would have allowed the company to use automated software, such as web scrapers and crawlers, to access and collect content from domains and URLs associated with business accounts to "develop, improve, research, train and maintain the integrity" of Meta products.
Critics, including the News/Media Alliance and legal scholars, warned that this broad wording could permit Meta to scrape news websites for AI training without paying for expensive licensing agreements. The policy would have applied to any business using Facebook, Instagram, WhatsApp, or Messenger for commercial purposes, including managing pages, running ads, or listing products.
In response to the backlash, Meta has removed the specific language regarding AI training from its upcoming commercial terms. While the company has narrowed the stated purposes of web crawling to advertising and business products, legal experts note that the revised policy still authorizes automated website crawling. Meta maintains that these updates clarify existing permissions and do not change how the company uses web crawlers, asserting that the changes are intended to ensure accurate product information for recommendations rather than scraping news sites.
Entities
Andy Stone · Ars Technica · Facebook · Instagram · Meta · Muse · News/Media Alliance · Nieman Journalism Lab · Nieman Lab · WhatsApp
Claims
What the coverage asserts, and how many sources carry each claim.
- [● 18 SOURCES] There is no opt-out mechanism for business accounts; continued use of Meta's services after the effective date constitutes acceptance of the terms. cryptobriefing.com · www.newsbytesapp.com · techround.co.uk · uk.pcmag.com · www.europesays.com · +13 more
- [● 26 SOURCES] Meta's updated commercial terms allow the company to use automated software to scrape content from domains and URLs associated with business accounts. cryptobriefing.com · www.newsbytesapp.com · techround.co.uk · uk.pcmag.com · www.europesays.com · +21 more
- [● 19 SOURCES] The updated terms are scheduled to take effect on October 30, 2026. cryptobriefing.com · www.newsbytesapp.com · techround.co.uk · uk.pcmag.com · www.theregister.com · +14 more
- [● 24 SOURCES] The new language appears to let Meta crawl and scrape websites associated with business accounts, including online publishers. cryptobriefing.com · www.newsbytesapp.com · uk.pcmag.com · www.europesays.com · www.zdnet.com · +19 more
- [● 9 SOURCES] The provision could allow Meta to access content that would otherwise require expensive licensing agreements. www.newsbytesapp.com · www.europesays.com · au.pcmag.com · entrevue.fr · www.upday.com · +4 more
- [● 3 SOURCES] The terms allow Meta to reach outside its walled gardens to websites associated with products to power AI models. uk.pcmag.com · www.europesays.com · au.pcmag.com
- [● 9 SOURCES] Meta claims the update is intended to ensure accurate product information for recommendations, not specifically to scrape news sites. www.searchenginejournal.com · www.europesays.com · tbreak.com · www.tokenpost.com · virasorovirtual.com · +4 more
- [● 14 SOURCES] Meta intends to use scraped content to develop, improve, and train its AI products. mercado.etc.br · au.pcmag.com · cryptobriefing.com · www.newsbytesapp.com · www.zdnet.com · +9 more
- [○ 1 SOURCE] Meta spokesperson Andy Stoner stated the update is intended to ensure accurate information about products for sale on third-party websites and is not about scraping news sites. www.searchenginejournal.com