One widely discussed advertising incident involved Meta’s Advantage+ tooling reportedly replacing a top-performing apparel ad creative with an odd AI-generated image of an elderly woman. The case matters because it moves the AI risk from production into distribution and optimisation.
For marketers, automated ad systems are attractive because they promise faster testing and performance gains. But when the system changes creative assets in ways the brand did not intend, optimisation can conflict with brand control.
What went wrong
The failure was not simply a strange image. It was a governance failure around who gets to approve creative changes. When an advertising platform can generate or alter creative assets automatically, the brand needs clear controls over what can be changed, when, and by whom.
Marketing lesson: AI optimisation should never override brand suitability without a human approval gate.
Governance questions
Which automated creative features are enabled? Can the team audit every asset variation shown to users? Is there a clear rule for when platform-generated creative must be manually approved before launch?
Source links: Business Insider, Fraud Blocker.