Why am I Still Seeing This: Measuring the Effectiveness Of Ad Controls and Explanations in AI-Mediated Ad Targeting Systems

Ad controls may not work as advertised: Meta's "See less" feature failed to significantly reduce ads on the topics users selected (Body Weight Control and Parenting), suggesting the transparency tool does not deliver on its promise to give users meaningful control over their ad experience.

Effectiveness varies by user demographics: The study found that the "See less" control was "less effective for some users whose demographics are correlated with the topic," indicating that certain groups may have even less ability to manage their ad exposure than others.

Explanations lack transparency: The majority of ad targeting explanations for local ads "made no reference to location-specific targeting criteria," meaning users were not receiving the information Meta markets as helping them understand why they see particular ads.

AI-mediated targeting reduces user agency: The shift from advertiser-specified targeting to AI-driven systems appears to have made it harder for users to control their ad experience, even when Meta provides tools ostensibly designed for this purpose, raising questions about whether AI optimization prioritizes advertiser goals over user preferences.

Transparency tools may create false sense of control: While Meta promotes ad explanations as a mechanism for users to "understand why they saw certain ads and inform actions to control future ads," the study suggests these explanations are often uninformative, potentially misleading users about their actual ability to manage their ad targeting.