Meta AI changing your ads without permission: who is liable after October 30?

What Meta changed

Meta is updating its Commercial Terms with a new "AI Features" section that takes effect on October 30, 2026. According to ADWEEK's reading of the terms, advertisers are "solely responsible" for everything Meta's AI creates or changes in their ads, including features Meta switches on itself. Before this change, that responsibility applied mainly when the advertiser chose to use Meta's AI tools. The terms also say AI features "may make mistakes and may not always produce accurate, complete, or suitable results."

What is already happening to advertisers

ADWEEK reports that Bookshop.org posted on Threads that an automatic account update had enrolled it in a generative AI feature. Meta then added an AI-generated illustration to a human-made ad without the brand's knowledge. Bookshop.org has since turned off all of Meta's AI features in ads.

The agency Exverus describes a wider pattern. In its account, Meta's AI adds sitelinks, translates copy, and creates AI images and carousels from a brand's own photos. Some of these changes appeared on ads where the advertiser had not enabled the feature. Exverus has pulled live ads that contained changes which got past its own quality checks. The article gives no numbers on how many ads are affected, so treat this as a warning sign, not a measured rate.

What a media buyer should do before October 30

1. Turn off the AI features you did not enable. In Ads Manager, go through Advantage+ options at the campaign and ad set level. Check the collapsed menus, because the article says many options are hidden there.
2. Audit live ads. Look at each active ad for translated text, extra sitelinks, AI images, or carousels built from your photos that you did not publish.
3. Read the "AI Features" section of the Commercial Terms and confirm whether it applies to your account. ADWEEK did not verify whether the terms cover every market.
4. Archive your originals. Save the source files, approved copy, and dated screenshots of live ads. If a platform-made change is disputed, this is your proof of what you published.

Our take: the risk is not only a changed image. An AI-generated variation that a brand never approved can put claims, prices or brand voice in front of buyers. If you run ads at scale, a weekly review of creative diffs is cheap insurance. Our TikTok post on automatic AI ads describes the same pattern on another platform, and the Meta product browsing opt-out guide covers another Meta default you should check before the end of October.

What is not confirmed yet

• Whether the liability clause applies to advertisers in every country, including Ukraine. The article does not say.
• Whether Meta makes advertisers liable for AI features they never switched on. Meta did not answer that question directly. Its spokesperson said the change folds existing generative AI ad terms into the main terms.
• The full text of the new terms. The article quotes only excerpts.
• The exact date of the Bookshop.org post.