Meta drops AI training clause from business terms, keeps crawling rights for linked sites
Meta removed language permitting AI model training on indexed website content from its commercial terms on October 17, following widespread criticism. The revised terms, effective October 30, still allow the company to crawl domains associated with business accounts.
Removal of AI Training Provisions
Meta deleted provisions on October 17 that authorized training AI models using content from indexed publisher websites, according to Unite.AI (Meta zmienia klauzulę dotyczącą indeksowania stron internetowych w projektowanych Warunkach handlowych). The updated terms no longer reference using such content for model training, Daily Sabah reported (Meta revises terms amid concerns over AI training on news content). UA.NEWS, citing The Verge, noted that Meta also struck a passage allowing the use of site content for research and AI product development (Meta usunęła z warunków sporne sformułowanie dotyczące treści stron — The Verge).
Continued Access to Business-Linked Domains
The amended terms preserve Meta's ability to scrape domains tied to business activities for advertising systems and commercial tools, Unite.AI reported (Meta zmienia klauzulę dotyczącą indeksowania stron internetowych w projektowanych Warunkach handlowych). Daily Sabah added that the company maintains rights to deploy bots and crawlers, and to generate test login credentials on pages connected to business accounts (Meta revises terms amid concerns over AI training on news content). Promptowy stated that each business account must consent to scraping of domains linked to its profile (Nowe warunki Meta mogą pozwolić jej trenować AI na treściach wydawców bez opłat). These provisions apply specifically to business accounts [[2], [3]].
Robots.txt Compliance Disputed
Legal experts consulted by Promptowy suggested the clause might override robots.txt instructions, potentially enabling Meta to train AI models on publisher content without compensation (Nowe warunki Meta mogą pozwolić jej trenować AI na treściach wydawców bez opłat). Meta spokesperson Daniel Roberts countered that the changes restrict crawling to advertising systems and do not modify the company's respect for robots.txt directives, UA.NEWS reported, citing The Verge (Meta usunęła z warunków sporne sformułowanie dotyczące treści stron — The Verge). PCMag noted a company spokesperson confirmed the updated terms grant no new powers that would supersede robots.txt preferences (Meta Sparks Alarm About It Scraping Websites for Potential AI Training).
Implementation and Consequences
The revised terms take effect October 30, 2026, for business accounts [[2], [3]]. AIHere reported that businesses and news publishers failing to meet these requirements will lose access to profile management on Meta's platforms (Nowe warunki Meta pozwalają jej AI przeszukiwać powiązane strony internetowe).
Sources
- [1]Meta zmienia klauzulę dotyczącą indeksowania stron internetowych w projektowanych Warunkach handlowych· Unite.AI
- [2]Meta revises terms amid concerns over AI training on news content· Daily Sabah
- [3]Nowe warunki Meta mogą pozwolić jej trenować AI na treściach wydawców bez opłat· Promptowy
- [4]Meta usunęła z warunków sporne sformułowanie dotyczące treści stron — The Verge· UA.NEWS
- [5]Meta Sparks Alarm About It Scraping Websites for Potential AI Training· PCMag
- [6]Nowe warunki Meta pozwalają jej AI przeszukiwać powiązane strony internetowe· AIHere
- #Meta
- #AITraining
- #BusinessTerms
- #WebCrawling
- #RobotsTxt
- #Publishers
- #AdvertisingSystems