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.

El fin de la privacidad, Dr Michal Kosinski

536 views

Published on

Presentación del Dr. Michal Kosinski, del evento del día 4 de julio Inteligencia Artificial "El fin de la privacidad, sus efectos en Gobiernos, Negocios y Personas"

Published in: Data & Analytics
  • Login to see the comments

El fin de la privacidad, Dr Michal Kosinski

  1. 1. Michal Kosinski | michalk@stanford.edu
  2. 2. 500MB daily data output per person, globally, in 2012 IBM
  3. 3. daily data output per person, globally, in 2025 The Economist, 2017 62 GB
  4. 4. What can we learn from digital footprints?
  5. 5. YouTube
  6. 6. How does the prediction model work?
  7. 7. How does the prediction model work?
  8. 8. Source: https://www.facebook.com/ads/audience-insights/
  9. 9. Predicting Personality from Facebook Likes
  10. 10. Work colleague (.27) Friend | Cohabitant (.44) Family member (.50) Spouse (.58) Average number of likes (227) Accuracy(correlation) Youyou, Kosinski, Stillwell (2015). Computer-based personality judgments are more accurate than those made by humans. Proceedings Of The National Academy Of Sciences.
  11. 11. Beyond Facebook Likes
  12. 12. Park, Schwartz, Eichstaedt, Kern, Kosinski, …, Seligman (2015). Automatic personality assessment through social media language. JPSP, 108(6), 934–952. High Openness -> <- Low Openness
  13. 13. Source: https://www.ibm.com/watson/developercloud/personality-insights.html
  14. 14. Ferry, Q., Steinberg, J., Webber, C., FitzPatrick, D. R., Ponting, C. P., Zisserman, A., & Nellåker, C. (2014). Diagnostically relevant facial gestalt information from ordinary photos. eLife, 2014(3).
  15. 15. Samochowiec, J., Wanke, M., & Fiedler, K. (2010). Political ideology at face value. SPPS, 1(3), 206–213.
  16. 16. Good proxy for: • Genetic factors • Hormonal factors • Developmental/environmental factors • Cultural factors & SES
  17. 17. Kosinski , Wang (in prep). Predicting Personality From Faces.
  18. 18. OPENNESS CONSCIENTIOUS NESS EXTROVERSIONAGREEABLENESS NEUROTICISM Personality profile
  19. 19. AI “dreaming” about the faces of introverts and extroverts Most Introverted Most Extroverted
  20. 20. Source: https://www.microsoft.com/cognitive-services/en-us/emotion-api
  21. 21. The End of Privacy
  22. 22. AI ‘dreaming’ about faces of homosexual and heterosexual males
  23. 23. Sexual orientation: Classification accuracy vs. the number of facial images per person.
  24. 24. Thank you! michalk@stanford.edu

×