Panel Discussion : Ethics in Data Science
summary
Panelists cite concrete failures and edge cases: a word‑association corpus Small World of Words with ~90,000 participants and ~1,000,000 responses carries entrenched sexist and cultural biases that corrupt downstream classifiers; public scraping plus geotags enables de‑anonymization of sex‑work networks; lack of consent in many hosted datasets clashes with GDPR obligations; corporate moves like the dissolution of Google's ethics board and the Facebook emotional experiment show governance and consent gaps that require technical and policy fixes.
AI-written summary · about
chapters (7)
- 0:00Introduction to ethics discussion
- 1:50Panelist introductions and concerns
- 8:14Bias and data origins in AI
- 13:05Trust and industry ethics challenges
- 23:44Public data use and boundaries
- 27:20Consent and data ownership
- 29:34Regulation and community responsibility
