Keynote: Digital Psychometrics and it's Future Effects on Technology
summary
Claims that psychometric models built from Facebook and phone data can predict Big Five traits and intelligence at 0.5 to 0.6 correlation, trained on a 7 million user database with about 3.5 million consented profiles via the Apply Magic Sauce API. Public reports allege use on 220 million US citizens with 175,000 ad variants. Validation shows ~10 likes beat a coworker, ~65 beat a friend, ~250 beat a spouse; an A/B campaign yields 6 million impressions, ~10,000 clicks and ~400 purchases, and phone sensors flag depression risk.
AI-written summary · about
chapters (21)
- 0:00Intro
- 0:45New stories
- 1:57What is psychometric
- 3:12Digital metrics
- 3:45Personal websites
- 4:14Apply Magic Sauce
- 4:54Predict My Profile
- 6:35How does it work
- 7:34The dataset
- 9:25The word cloud
- 10:27The word agreeableness
- 12:02The latest development
- 13:04How accurate is it
- 15:30Question
- 15:53The Trump Story
- 17:17Healthcare
- 18:10Facebook
- 19:28Future of Digital Psychometrics
- 20:24Lucy Kay
- 21:05Most of our life
- 21:26Conclusion
related talks
05mInterview with Alex Willcock
24mHow to Target UX to Different Psychographics
25mUnveiling and Quantifying Facebook Exploitation of Sensitive Personal Data for Advertising Purposes
43mKeynote: Consequences of an Insightful Algorithm
21m“Trust Us”: Mobile Phone Use Patterns Can Predict Individual Trust Propensity
26mThe Need for People Based Measurement
19mSocial Media - Did We Reach the Summit? pt.1
30mWhen worlds collide