Meta is changing the commercial terms that govern how businesses use its products, with the new wording due to take effect on 30 October. The changes, first reported by CIO.com, move responsibility for errors made by Meta’s AI features onto its business customers, widen the licence Meta takes over content shared on its platforms, and set out new rights for the company to crawl websites linked to business accounts.
Under the proposed terms, Meta accepts that its AI features can make mistakes and produce inaccurate or unsuitable results, but states that customers are “solely responsible for and assume any and all risks” linked to those features and the content they generate. Frank Dickson of Dickson Research told CIO the clause turns accountability on its head, since Meta builds the tools and admits they can fail. Mike Wilkes of Aikido Security pointed out that only Meta controls the models producing the output, and warned that terms like these could drive enterprise customers away.
The terms also give Meta a licence to use anything businesses share, post or upload across its platforms. This includes material placed by third parties acting for a business, such as an agency running ads or a contractor managing a Facebook page. Meta may use that content to provide, develop, improve, research and maintain its products, and the list of products covered sits on a separate page that Meta can update at any time.
The changes Meta made in July already describe that licence as perpetual and that it survives the end of the agreement. Noah Kenney of Digital 520 said this means the licence only ends when content is deleted from Meta’s systems, so walking away from the platforms would not end Meta’s rights over what has already been shared. Meta has also recently removed explicit reference to training its models on business data, though Thomas Randall of Info-Tech said the revised wording is still broad enough to cover most AI development. Continued use of Meta’s products counts as acceptance of the new terms.
The section drawing most attention from publishers covers web crawling. The draft allows Meta to use scrapers, bots and crawlers to access, index, cache, and collect content from domains and web pages associated with a business’s products or services, and to pull updates from them periodically. Aman Mahapatra of Tribeca Softech said the wording ends a long held assumption that robots.txt protects published content.
Meta has since told CIO that the crawling terms apply only to its ads and business products, that site owners can still use robots.txt to set their preferences, and that the update is not about scraping news sites. The company said the published wording would be amended to reflect this, although it had not changed at the time of reporting.
For Independent Media Association members who use Facebook, Instagram, WhatsApp and Threads to reach audiences, the changes are worth reviewing before 30th October. Anything posted to Meta platforms, including by freelancers or agencies on a publisher’s behalf, falls under the expanded licence, and a publisher using Meta’s AI tools to create posts or ads would carry the legal risk if the output turns out to be wrong. Members should also watch whether Meta’s promised clarification on crawling appears in the final terms.