Anti-Privacy Anti-Patterns
summary
Defines privacy as user control over both tangible PII and derived behavioral data, and argues web search and personalization plus cheaper servers lead companies to collect everything. Lists six anti‑patterns: inference without consent, lack of transparency, erosion of control, blind trust in black‑box ML (illustrated by Microsoft’s Tay), an “internet of integrations” with apps requesting broad Dropbox‑style permissions, and unclear data ownership (Cambridge Analytica: 270k quiz users, ~87M affected). Recommends explicit consent, fewer opaque ML uses, and local processing where possible.
AI-written summary · about
chapters (9)
- 0:00Introduction and speaker intro
- 3:08Defining privacy and assumptions
- 9:33History of search and personalization
- 15:12Inference without explicit consent
- 18:48Lack of transparency and erosion of control
- 20:10Trust in black box algorithms
- 21:35Internet of integrations and unclear data ownership
- 23:00Recap of anti-patterns and solutions
- 24:23Q&A session
related talks
47mDesigning for User-Centered Privacy
11mThe Future of Privacy
33mDesigning and Building With Privacy In Mind
37mThe balance between data privacy and user experience
43mPanel: Privacy Design Patterns for AI Systems: Threats and Protections
13mFuture Trends in Privacy Preserving Technologies
22mPrivacy Threat Modeling
28mPrivacy Engineering Networking Break AMA