Meta added a new “AI Features” section to its Commercial Terms on Oct. 30, making advertisers solely responsible for any AI-generated edits to their ads. The change follows incidents where Meta’s AI automatically altered ads without consent, creating new compliance and brand-safety challenges for marketers.
What happened
Meta’s revised Commercial Terms, effective Oct. 30, introduced an “AI Features” clause that says advertisers are “solely responsible” for any changes Meta’s AI makes to their ads, even when the AI is turned on automatically. The terms also note that the AI “may make mistakes and may not always produce accurate, complete, or suitable results.” No separate contract was required; continued use of the platform after notice counts as acceptance.
The shift came after several brands reported unwanted AI alterations. Bookshop.org, an online bookseller, discovered that a generative-AI feature added an illustration to one of its ads without consent. The bookseller wrote on Threads that the ad was “enhanced” (or ruined, really) by Meta without their knowledge and subsequently disabled all AI ad features.
Meta’s engineers have already admitted that one of their AI agents began hallucinating in production, prompting a redesign that split the agent into 14 specialized sub-agents. About two-thirds of the system’s changes now land without human edits, meaning roughly a third still need correction.
Why it matters
Shifting 100% of legal liability onto advertisers creates operational risk. Agencies such as Exverus reported that Meta’s AI automatically added sitelinks, translated copy, and generated images from brand assets, sometimes on campaigns where the features were never enabled. Ryan Adams of Exverus warned that hidden menu options can apply enhancements unless manually unchecked.
Legal analysts note that Meta’s liability is still capped at the greater of $100 or the fees paid in the prior twelve months, and advertisers must indemnify Meta against third-party claims. However, the new terms remove the previous condition that liability only applied when advertisers opted into AI tools, expanding risk exposure for any brand running ads on Meta.
Reputational risk is also a concern. Asa Hiken of eMarketer said consumers are unaware of Meta’s ad policies and may assume the brand deliberately used AI, potentially damaging trust.
What this means for marketers
- Audit every Meta campaign immediately and turn off all generative-AI features in Ads Manager.
- Verify that no hidden settings are auto-applying enhancements; manually uncheck any collapsed options.
- Update compliance checklists to treat AI-generated creative as a separate risk category.
- Consider a verified lead generation service that screens ad content for compliance before it goes live.
- For brands focused on search visibility, ADFLU’s autonomous SEO service can help ensure AI-driven ad changes do not affect organic rankings or cause unwanted citations in AI answers.
- Legal teams should revise contracts to reflect the new liability language and negotiate indemnity clauses where possible.
- Deploy monitoring tools that flag AI-generated assets in real time to reduce accidental brand-unsafe content.
"This particular ad was 'enhanced' (or ruined, really) by Meta without our knowledge" Bookshop.org (bookseller)
The bottom line
Meta’s policy shift places full responsibility for AI-edited ads on advertisers, increasing both operational and reputational risk. Brands should disable default AI features, tighten compliance processes, and consider third-party verification to protect their campaigns and maintain brand safety.
Sources
- 1.ppc.land
