Machine Learning Ethics
summary
Examples show real harms and fixes: Amazon hiring model favored men and ad targeting hid executive roles from women, an A‑level algorithm penalises low‑income areas, facial recognition misidentifies non-white faces, and an Apple Card decision gives a spouse a 20× lower limit. Causes include unrepresentative training data, preprocessing steps that introduce bias, and different algorithms producing different subgroup errors. Practical steps are break down accuracy by group, check whether addresses infer ethnicity, start in back‑office and hold model owners accountable.
AI-written summary · about
chapters (8)
- 0:00Introduction to AI ethics
- 1:14The role of AI in industry
- 3:09Real-world examples of bias
- 5:49Defining fairness in models
- 9:30Sources of algorithmic bias
- 12:39Importance of ethical AI
- 15:27Practical tools and tips
- 19:08Final thoughts on bias
related talks
50mEthical Machine Learning
35mBuilding Inclusive Machine Learning Models
26m(Un)ethical Artificial Intelligence: How to Keep the AI Fair for Everyone
47mEthical Machine Learning
54mEthical Machine Learning
34m(Un)Ethical Machines? Why AI Bias Is a Problem and What to Do About It
47mEthics & AI: Identifying & preventing bias in predictive models
56mAn AI with an Agenda: How Our Cognitive Biases Leak Into Machine Learning