Meta has indefinitely paused planned updates to four business agreements — including its Ad Creative Generative AI Terms — after weeks of backlash from publishers, advertisers, and regulators over provisions that would have allowed the company to scrape content from websites linked to Facebook, Instagram, WhatsApp, and Threads business accounts for use in its products, including generative AI systems.
The changes were originally scheduled to take effect on October 30, then postponed to November 20 following initial criticism. On November 5, Meta announced the updates were paused entirely, with no new effective date set. The company said it would continue to engage with stakeholders while its existing terms remain in force.
What the Terms Would Have Done
The four affected agreements — Commercial Terms, Product Catalog Terms, Self-Serve Ad Terms, and Ad Creative Generative AI Terms — contained language that critics argued gave Meta sweeping rights to data published on any website a business links from its Meta accounts. Digital publishers and trade groups warned the terms could allow Meta to train its AI models on proprietary content without compensation or consent.
Meta attempted to defuse the backlash by removing the word "train" from the proposed terms and narrowing the stated use of scraped content to advertising and business products. The revisions did little to satisfy critics. Courts in Norway and Brazil issued preliminary injunctions sought by media organizations, and the UK Competition and Markets Authority opened a formal investigation into whether Meta was imposing exploitative terms on publishers and advertisers.
Why It Matters for Companion and Creative AI Developers
For developers building AI companion apps, NSFW AI tools, or creative AI services that maintain Meta business accounts or advertise on Facebook and Instagram, the paused terms would have granted Meta broad rights to scrape content from linked websites. A companion app's marketing site, privacy policy page, or feature documentation linked from a Meta business profile could have been fed into Meta's training pipelines without the developer's explicit consent.
The Ad Creative Generative AI Terms are particularly relevant to the AI companion space. These terms govern how AI-generated content is handled within Meta's advertising ecosystem — the same ecosystem that many companion and NSFW AI developers rely on for user acquisition. Had the terms gone into effect, Meta could have used AI-generated ad creatives and linked business content to further train its generative models, raising questions about intellectual property and data rights for developers operating in the adult AI space.
Meta's ownership of Muse, its AI companion app, adds another dimension. The same company that operates a direct competitor in the companion space would have gained access to web properties linked from business accounts used by rival companion developers. While Meta has stated the scraped data would be used for advertising and business products, the breadth of the proposed terms left little room for opt-out.
Regulatory Pressure Mounts
The backlash wasn't limited to public criticism. Legal challenges emerged on multiple fronts. Norwegian and Brazilian courts issued injunctions blocking the terms, and the UK's competition watchdog opened a formal probe. Ars Technica noted that Meta's attempt to quietly push through the terms backfired, with the headline asking whether Meta thinks users "can't read" — reflecting the widespread view that the terms were deliberately opaque.
Meta's statement on the pause was brief: the company said it was "pausing the changes as we continue to engage with stakeholders." The company has not indicated whether the terms will be revised and resubmitted or abandoned entirely.
What's Next
Meta's existing Commercial Terms remain in effect. For companion and NSFW AI developers, the pause is a temporary reprieve, not a permanent resolution. The episode signals that Meta is actively seeking broader data access for generative AI training, and similar proposals are likely to resurface. Developers who rely on Meta's platforms for user acquisition or business operations should watch for updates and consider how their linked web properties could be affected by future terms.