Successfully reported this slideshow.
We use your LinkedIn profile and activity data to personalize ads and to show you more relevant ads. You can change your ad preferences anytime.

A Privacy Framework for Social Machines

118 views

Published on

SOCIAM all-hands, September, University of Oxford

Published in: Technology
  • Be the first to comment

  • Be the first to like this

A Privacy Framework for Social Machines

  1. 1. A Privacy Framework for Social Machines Kieron O’Hara
  2. 2. Motivation 1: ‘A Concept in Disarray’ • “Any attempt to locate a common denominator for all the manifold things that fall under the rubric of ‘privacy’ faces an onerous choice. A common denominator broad enough to encompass nearly everything involving privacy risks being overinclusive or too vague. A narrower common denominator risks being too restrictive.” Solove, Understanding Privacy 2
  3. 3. Motivation 2: Social Machines Are All Shapes and Sizes
  4. 4. Taming Complexity • Principles for a framework to make sense of the disarray • Separate out loose hierarchy of privacy discourses • Help defuse and organise privacy debates • Help manage privacy issues for SMs • Separate out: • Values/ethics • Legal issues • Cybersecurity
  5. 5. Level 1: Conception • What conceptions of privacy are relevant to the social machine? • Does it deal with personal data? • Does it use names/pseudonyms/anonymity? • Location? • Non-digital aspects (e.g. F2F)? 5
  6. 6. Level 2: Actuality • Is there a breach of each conception of privacy? • A matter of fact & measurement • It does NOT mean • Have my rights been breached? • Has the law been broken? • Have my interests been harmed? • Have I noticed anything untoward? • Has anything happened that I care about? • Cybersecurity/implementation 6
  7. 7. • What does the breach/non-breach feel like? • Shame, outrage, creepiness, pride • Do I even notice? • Little work at this level Level 3: Phenomenology 7
  8. 8. 8
  9. 9. Level 4: Preferences • What do I want? • When do I want visibility to the SM? • When do I want to be concealed? • What do others want of me? • What exposure to others do I want? • Idiosyncrasy rules • Control, consent, privacy markets 9
  10. 10. Level 5: Norms • Regularities, conventions, expectations • Variations across culture, classes • What do participants expect of an SM? • How do norms carry over into SMs? • Relation to other norms • Can be used to derive rules 10 https://wordspictureshumor.wordpress.com/2013/03/06/blind-humor/
  11. 11. Level 6: Law & Regulation • Privacy is not a legal concept! • Unlike data protection • Regulated but not created by law • Organisational rules • Privacy law • What does the state/the organisation constrain? 11
  12. 12. Level 7: Politics & Morality • What is right/wrong? • Value • Political effects • Democracy (Westin) • Security (Etzioni) • Autonomy of the citizen (Rössler) • Social interests outside the SM 12 https://popularresistance.org/developing-a-people-centered-security-culture/
  13. 13. Questions for Social Machines • 1: What conceptions of privacy are implicated? • 2: Is privacy protected? • 3: Does the design convey (lack of) protection? • 4: How do participants exert control? • 5: What are participants’ expectations? • 6: Who is accountable for data breach? • 7: What is the value of the SM?
  14. 14. SOCIAM Across the Levels • 1: group privacy • 2: privacy-preserving ML, ethical data initiative, transparency/output • 3: transparency/awareness • 4: transparency/recommendations (e.g. X- ray Refine), PDSs, data terms of use • 5: privacy of children’s data • 6: web observatory rules • 7: ethics
  15. 15. Across the Levels • Data Safe Havens • 7: Public good: medical research v privacy • 6: What sharing can we legally do? • 6: What organisational rules do data controllers have? • 6: What protocols can we craft to govern a federation of organisations? • 5: Need to preserve public confidence in medical data sharing • 4: (Future work:) machine-readable data terms of use • 2: Automation of experiment design
  16. 16. Conclusion • 7-level privacy framework • Disentangle separate issues in debate • Organisational principle for SM management • Loose hierarchy of discourses/issues to resolve

×