Meta’s ad systems allowed promotion of an app that promised to create nude images of female politicians, according to Ars Technica.
The report says one ad featured pornographic video with a deepfake closely resembling a U.S. politician. The episode points to a recurring failure mode for generative AI platforms and ad networks: tools that can produce sexualized nonconsensual imagery can be packaged and marketed faster than moderation systems catch them.
The practical harm is not abstract. Deepfake abuse can target public figures, private individuals, and political opponents, while paid ads can amplify the tools that enable it. Platform review failures also make enforcement look reactive rather than preventative.
The article focuses on Meta’s ads and the app’s promotion, not a complete audit of every platform hosting similar material. Still, it is a concrete example of why AI image abuse is as much an advertising and distribution problem as it is a model-safety problem.