Meta's AI Ad Tools Are Generating Brand Disasters — And Meta Says It's Advertisers' Problem
Summary
- • Meta's AI ad tools are producing distorted products, gibberish text, and inappropriate imagery — eight advertisers and agency execs tell Business Insider this has become routine
- • A recurring bug auto-toggles AI creative settings on for advertisers who have disabled them, affecting most clients at one major agency managing hundreds of simultaneous Meta campaigns
- • REI drew consumer backlash after Meta 'auto-enrolled' it in an AI feature that generated a nonsensical double-handlebar bike ad without the brand's knowledge
- • Meta's official stance: its terms of service state AI 'can make mistakes' and advertisers bear full responsibility for reviewing all AI-generated outputs
Details
Meta AI ad quality failures now routine
Eight advertisers and agency execs confirmed to Business Insider that distorted product imagery, garbled text, and AI-invented product variations are now routine outcomes from Meta AI ad tools — not isolated incidents.
Auto-enrollment bug persists
A recurring bug accidentally toggles AI creative settings on for advertisers who have disabled them; one major agency reports it affects most of its 15 Meta clients and flagged the bug to its Meta rep as recently as the week of publication.
REI public backlash incident
Outdoor retailer REI faced consumer backlash after Meta 'auto-enrolled' it in an AI feature that generated an Instagram ad showing a bicycle with two handlebars. REI said the image was 'inaccurate' and 'inappropriate' and that it had not consented to the AI feature.
Meta shifts all liability to advertisers
Meta's terms of service state 'AI can make mistakes and that it is the advertiser's responsibility to review the AI outputs' — leaving brands legally and reputationally exposed for errors generated by Meta's own AI systems.
Major ad budgets exposed
Flat Circle agency manages approximately $200M in annual Meta ad spending for direct-to-consumer brands; TAU Marketing manages approximately $500M annually across platforms. Both report brand safety concerns caused by Meta AI errors.
AI error-checking as new standard procedure
Advertisers now must manually verify AI settings for each campaign to catch inadvertent activations or altered outputs. With some agencies running hundreds or thousands of simultaneous ads, checking AI settings has become 'a new standard operating procedure,' per one agency CEO.
Specific AI error examples
A pajama brand had Meta AI swap a pajama dress for a shirt-and-pants set; a women's Montana networking group had men added to its ads; a book campaign ran with garbled product text and distorted product images for ~12 hours before the advertiser was alerted by friends.
Distilled from Business Insider reporting based on interviews with eight advertisers and agency executives, including named executives from Mediassociates, Flat Circle, and TAU Marketing.
What This Means
Meta's aggressive rollout of AI ad tools is creating a brand-safety crisis for advertisers, with persistent bugs auto-enabling AI features that generate inaccurate, distorted, or brand-damaging imagery. Meta's response — citing a terms-of-service clause that places full responsibility on advertisers — has failed to address the scale of the problem, particularly for agencies running hundreds of simultaneous campaigns where manual review of every AI setting is impractical. The burden falls hardest on smaller brands with limited resources to audit every ad, while Meta continues expanding AI features and declining accountability for errors its systems generate. This represents a genuine commercial risk for Meta if advertisers reduce spend on a platform that generates brand-damaging content without consent, and a broader signal that AI-generated ad creative is not yet reliable or transparent enough for autonomous deployment at scale.