Vision: Human-as-the-Unit Privacy Management with AI Agents
summary
Interviews with 12 users surface nine cross-context privacy challenges spanning apps, time, and relationships, and researchers generate nine agent concepts to address them. A speed-dating survey with 116 US participants ranks post-sharing privacy management highest; top designs are automated remediation tools such as History Swapper, which deletes outdated sensitive data like old addresses on Amazon. Users report greater confidence in AI than manual efforts but raise risks of overdelegation, single points of failure, and call for adjustable autonomy, data minimization, and robustness.
AI-written summary · about
related talks
11mHow Does AI Change Privacy? Towards Helping Practitioners Scope, Motivate, and Conduct
29mAnti-Privacy Anti-Patterns
43mNext in Privacy Innovation: AI Agents that Make Governance Scalable
13m"We are not Future-ready": Understanding AI Privacy Risks and Existing Mitigation...
15m"There will be less privacy, of course" How and why people in 10 countries expect AI...
21mSecurity in the Age of AI & Copilot
22mFireside Chat - AI Security Vulnerabilities
01h 17mPrivacy and New Technologies