Questions and Discussion
Talkat Wrong at the Root: Racial Bias and the Tension Between Numbers and Words in Non-Internet Data · 2019-06-05..06-07·
summary
Panelists argue for concrete outputs: a short list of foundational papers, shared reading resources, and new venues that reward domain-specific, reproducible work. They propose algorithmic tests that reject the premise when features are insufficient and emphasize supervised-learning limits. They cite practical failures—facial-genetics projects produce a dozen cases since 2015 with no exact matches and generated faces that humans often misidentify—and call for explicit norms, pedagogy, and cross-field audits.
AI-written summary · about
chapters (13)
- 0:00Intro
- 2:15How can we support you
- 3:07Shared resources
- 4:19Working groups
- 6:48Informed critique
- 7:28Supervised learning
- 11:32Work in progress
- 13:19Political ideology
- 15:25Fairness is political
- 20:29Collapse of Civilizations
- 23:29Uncover or Discover
- 24:31Algorithms
- 26:39Discussion
