Societal Concerns in Targeted Advertising
summary
Three concrete failures of targeted ads and steps to fix them: custom audiences and Audience Insights let an uploader infer a single person’s attributes (reported and patched via a bug bounty); location inclusion/exclusion composes to Planned Parenthood microtargeting so ads can hit people inside a clinic; ad delivery optimization produces a persistent skew so identical job ads get different audiences — racial skew in ad delivery (72% vs 30% white) — even when targeting 50/50, and Facebook’s policy changes (removing age/gender/ZIP for housing/employment/credit) only partly address the problem.
AI-written summary · about
chapters (20)
- 0:00Introduction
- 1:02What is Targeted Advertising
- 1:58Targeting
- 4:15Custom Audiences
- 5:20Lookalike Audiences
- 17:55Audience Insights
- 19:52Custom Audiences Audience Insights
- 27:37Targeting by Location
- 29:33Location Controls
- 30:18Automation of Hate
- 33:25Logistics
- 36:58Presence
- 38:53Conceptualisation
- 40:39Unbaked Ideas
- 41:45Logistics Operation
- 44:32Hiring the right people
- 45:27Discrimination and unfairness
- 47:27Facebooks response
- 49:14The delivery is a complex process
- 52:03Measuring racial bias
related talks
22mExploring User Perceptions of Discrimination in Online Targeted Advertising
01h 24mfrom “old” filter bubbles to political microtargeting.
32mPlatform-supported Auditing Of Social Media Algorithms For Public Interest
01h 07mPolitical behavioural targeting
01h 07mThe great data race ...
28mAdtech Privacy Risk: Increasing Regulation and Enforcement
44mFireside chat: Solving media and advertising challenges
25mUnveiling and Quantifying Facebook Exploitation of Sensitive Personal Data for Advertising Purposes