Adweek reports that advertisers have started pulling live Meta ads after finding AI-generated changes they did not ask for, from added illustrations to translated copy. The timing matters: from October 30, Meta's updated Commercial Terms make advertisers solely responsible for what its AI features create or change. Agencies are responding by switching off automation by default.
What advertisers found in their ads
Bookshop.org said an automated account update opted it into a generative AI feature, after which Meta added an AI-generated illustration to an ad its team had made by hand. The company posted about it on Threads and has since turned off all Meta AI ad features.
Digital agency Exverus told Adweek that Meta's AI added sitelinks, translated copy, and built AI images and carousels from brands' own photos, in some cases on ads where nobody had enabled those options. Even after several rounds of quality checks, the agency pulled live ads to strip the changes out. Its Ryan Adams said many enhancements sit in hidden or collapsed menus and keep running unless each one is unchecked manually.
- AI-generated illustrations added to human-made creative
- Automatic translations of ad copy
- Sitelinks added without request
- AI images and carousels built from brand photos
Why October 30 raises the stakes
Meta's revised Commercial Terms add an "AI Features" section. Under it, the advertiser carries the risk for whatever Meta's AI produces or edits, including features Meta switches on itself. Previously that liability applied when an advertiser chose to use the tools. Meta's terms also warn its AI "may make mistakes" and may not produce suitable results.
Asked whether advertisers are liable for features they never enabled, a Meta spokesperson did not answer directly. Meta said individual Advantage+ features can be turned off in Ads Manager at any time and described the change as folding existing generative AI terms into its main ad terms.
Who this hits hardest
Regulated and brand-sensitive advertisers carry the most exposure: finance, health, and anyone whose claims are reviewed by legal before launch. An AI-rewritten headline or translated line that was never approved can create a compliance problem, and under the new terms it is the advertiser's problem. Agencies running many client accounts face the same issue at scale, because defaults can be re-enabled by account updates.
AdCellerant's Kayla Loecker told Adweek her firm turns off Meta's default AI automation on every campaign. Go Fish Digital's David Dweck said he won't tell clients to cut Q4 spend, but noted the opt-ins are landing right before Black Friday.
What to do this week
Open every active ad and expand all Advantage+ creative and enhancement menus, including collapsed ones, and uncheck anything you have not deliberately approved. Use the ad preview to compare what will actually serve against your original assets. Repeat the check after any account-level update, since that is how at least one advertiser was opted in.
If you do want AI enhancements, test them in a separate ad so you can review output and measure lift on its own. Keep screenshots of your settings at launch in case you later need to show what you approved.
What it means for you
- From Oct 30, you are responsible for whatever Meta's AI adds to or changes in your ads.
- Check collapsed enhancement menus on every live ad and uncheck anything you didn't approve.
- Re-audit after account updates, since defaults can be switched back on.
- Test AI enhancements in separate ads instead of letting them edit approved creative.
Sources
- Advertisers Pull Meta Ads That Had Unwanted AI Glow-ups (Adweek)
- Meta makes advertisers solely responsible for its AI ad edits from Oct 30 (PPC Land)
