Low rate loans for ladies, stags pay extra: The Role of Ethics in AI/ML
summary
Asserts that no single formalization of fairness satisfies all intuitions and shows this with concrete code and data not slogans. Demonstrations include the simplest algorithm: sort spreadsheet that produced Walmart style anti‑theft targeting, a linear regression simulation that sometimes corrects biased inputs and sometimes reinforces group gaps, replicated findings that women default far less in microcredit studies, and a concrete tradeoff where 75% of Black applicants are rejected at FICO 600 so every policy choice sacrifices some value.
AI-written summary · about
chapters (8)
- 0:00Start
- 6:19Individual Fairness
- 14:50Simple model: linear regression
- 17:36Does linear regression become biased?
- 24:19What if the input data is biased?
- 31:30Protected class is just another feature
- 36:57Exposing tradeoffs
- 40:01Uncomfortable Questions
related talks
26m(Un)ethical Artificial Intelligence: How to Keep the AI Fair for Everyone
40mResponsible AI Practices: Fairness in ML
35mOpen Source Fairness
11mAddressing Privacy, Bias, and Fairness in AI Systems
40mEthical AI: Addressing Bias & Algorithmic Fairness
42mFairness in Machine Learning: From Theory to Practice
35mBuilding Inclusive Machine Learning Models
23mMachine Learning Ethics