Meta has updated its commercial terms for business customers. According to reports by The Verge and Ars Technica on October 16, 2026, the new document could allow the company to use content from news publishers' websites for free to train artificial intelligence (AI) models. Ars Technica estimates the move could let Meta avoid millions of dollars in AI licensing fees owed to publishers.
What the Updated Terms Say
According to The Verge, the updated commercial terms could let Meta train its AI models for free on content from news publishers' websites. The document is a general set of terms applied to business customers, and The Verge reports that the provisions in question concern news publishers' websites specifically — that is, while the terms are general, their impact could first and foremost affect publishers.
"Meta's updated commercial terms could allow the company to train artificial intelligence models for free on content from news publishers' websites," writes The Verge (source).
Notably, commercial terms are not negotiated separately with each customer — they apply uniformly to all customers using the platform for business. A change to a single clause can therefore affect many companies at once. If this interpretation proves correct, Meta could use publishers' content to train its large language models without separate licensing agreements.
The Licensing Fee Question
According to Ars Technica, Meta could use the new, sweeping commercial terms to avoid paying millions of dollars in AI licensing fees to news publishers. Ars Technica described the new document as sweeping and characterized it as a step that could save the company millions of dollars (source).
Licensing fees have been one of the main legal ways AI companies use publishers' content: a company signs a deal with a publisher, uses the content to train models, and pays for it. According to Ars Technica, Meta may have found a way to avoid these fees in its general commercial terms — that is, to gain broad rights through a single document without negotiating separately with each publisher. This concerns AI licensing fees specifically — the fees paid for the right to use publishers' content to train models.
The December Deals: There Was Another Path
According to Reuters, in December 2025 Meta struck commercial deals on AI data with several major publishers: USA Today, People Inc, CNN, Fox News, The Daily Caller, Washington Examiner, and Le Monde. The deals allowed the company's AI chatbot to display real-time news alongside links to publishers' articles — that is, users received the latest news directly through the chatbot, with a link to the source (source). Reuters reported the deals based on an Axios story.
These deals show that a paid-partnership model worked in Meta's relationship with publishers: the company paid for content, and in return its AI chatbot delivered publishers' latest content to users with links. The new terms could become an alternative to exactly this model — free use.
So a year later, Meta could rely on general commercial terms instead of separate deals to use the same publishers' content — a sharp departure from its previous model of cooperation with publishers.
The Dispute Between Publishers and AI Companies Intensifies
The right to use content for training AI models has become one of the most contentious issues between publishers and technology companies. For example, on October 9, 2026, USA Today sued OpenAI, accusing it of copyright infringement in training artificial intelligence — our site previously reported on this. Thomson Reuters also won an appeals court victory against ROSS Intelligence — a decision described as one of the first major rulings on "fair use" in AI training.
Against this backdrop, Meta's new move draws particular attention: the question of where AI chatbots get news and who gets paid for it may now be settled not through individual contracts, but through platforms' general terms updated unilaterally. Amid such disputes, Meta's move could be an important signal for publishers: if a platform gains the right to use content for free through general terms, publishers' positions in future negotiations could weaken.
