Researchers have developed a method to infer sensitive user attributes from interactive targeted advertising systems. The study models the advertising channel as a noisy oracle, separating targeting predicates, exposure, interaction, and disclosure to capture the gap between eligibility and advertiser visibility. A reproducible benchmark was created using synthetic populations and a simulator to evaluate various inference attacks, finding that repeated campaigns with identity exposure yield measurable but bounded inference signals. The research highlights disclosure policy as the strongest control, with aggregate reporting and randomized disclosure significantly reducing the released signal. AI
RANK_REASON The cluster contains a single academic paper published on arXiv detailing a new research methodology and benchmark. [lever_c_demoted from research: ic=1 ai=1.0]
AI-generated summary · Google Gemini · from 1 sources. How we write summaries →