Academic sessions 1.
summary
Four concrete claims and failures: access governance works better when researchers are treated as partners so systems favor shared responsibility and self‑policing rather than a single “policeman” gatekeeper; ad hoc anonymization already fails in practice — the NYC taxi release of 173 million taxi rides had its hashing broken in an hour and shows auxiliary data makes reidentification persistent; formal methods like differential privacy and query‑only access can give provable bounds; EU cookie law (Article 53) flips to consent but industry resistance and weak regulator guidance leaves compliance as banners and tiny links, not effective protection.
AI-written summary · about
chapters (13)
- 0:00Start
- 3:47Summary
- 4:17data for research
- 5:45The default option policing
- 6:36The problem
- 7:48legitimacy
- 11:42Investing in people
- 12:19Evidence from data enclaves
- 13:36The 'Five Safes' model
- 25:02Erosion of Privacy
- 48:27The Privacy Diaspora
- 54:53The Three Vectors
- 58:32The behavioural advertising cookies' relationship
related talks
01h 13mAcademic sessions on data processing and privacy
01h 18mACADEMIC SESSION 3
01h 14mSenior Academic Session I.
01h 12mACADEMIC SESSION 1
01h 01mAcademic sessions on data protection law and privacy
01h 06mPanel on Interpretability in the Law
16mPrivacy by Design for JavaScript Data Systems
01h 14mSenior Academic Session I