Ethics Panel
summary
Concrete examples and experiments anchor the conversation: Google Translate’s gender fixes and a BERT probe that predicts “mathematician” as male ~83% (and grows more male with modifiers), a Vianai-style platform that exposes black box models by letting users perturb training data and surface uncertainty, and calls to test for robustness and accountability up front rather than after failures. The talk urges diversity, society in the loop, interdisciplinary review, and collective workplace action to refuse shipping harmful systems.
AI-written summary · about
chapters (11)
- 0:00Intro
- 8:00What is your background
- 9:49What is your experience in Africa
- 10:53What is your background in mathematics
- 16:48How do you recognize bias
- 20:07Being proactive rather than reactive
- 21:56Can we cater to everyone
- 22:32Fairness and equity
- 23:53Accountability
- 26:17Collective Action
- 35:06Hopes
related talks
56mEthics in Machine Learning Panel
33mEthics in Machine Learning Panel Q&A
31mPanel Discussion : Ethics in Data Science
47mEthics in AI Panel
38mSocial Implications of Bias in Machine Learning
31mCan you Build Algorithmic Justice?
55mPanel: Gender bias in AI - In collaboration with Women in Data Science Belgrade
37mKeynote: Ethical Machine Learning: Creating Fair Models in an Unjust World