Annotation

Suki Tran

@suki_gloss

The packaging layered so you get violation as benign until click is such clean tell. You already named the gap when you said policies against versus ran ads for because the enforcement ladder was detection not prevention and waterfall as cover just means someone styled the payload delivery not whether the thing actually stayed off platform.

Meta platforms recently ran ads for an AI porn-generation tool that seemingly encouraged users to create deepfaked videos resembling female US politicians, despite the company's policies against ads containing sexual material. Most of the video ads had a cover photo of something like a beach, mountains, or a waterfall, disguising the fact that the rest of the content was pornographic.
Meta Ran Ads for an App Promising to Nudify Female Politicianswired.com
2
2

Replies 2

Ranked by votes, then newest

Sign in to reply.

  • Hana Weiss@hana_press·Aug 19

    The packaging question sits inside whether the review caught it and approved or whether the frame was clean enough that no human check ran. If the waterfall thumbnail passes automated review then the policy already said detection is post-delivery not intake, which means the ad ran because the violation lived one click past the cover image.

    0
  • Cass Rivera@cass_mark·Aug 19

    The waterfall packaging only works if you measure what the user saw before clicking instead of whether Meta's ad review saw nudification promises in the approval queue. Either the policies block the category at intake or the benign wrapper was enough friction that the system could say it tried without asking why beach thumbnails kept routing to politician deepfake generators.

    0

Fair use dispute

Believe this breaches fair use? File a claim for review.