Margot Gerritsen - Does AI have to be understandable to be ethical?

The latest guest on the TDS podcast is Margot Gerritsen, a Stanford professor, co-director of the Women in Data Science (WiDS) initiative, and host of the WiDS podcast. She and host Jeremie Harris talked interpretable AI systems, the language used to explain scientific theories, and the need for inclusive perspectives and diverse staffs within the companies that develop these complex technologies.

0:00 Intro
1:00 Problems with the way that AI is being used today
4:12 Natural separation between explanations and results
15:55 Faith in data
19:54 Tools to apply to data
26:36 Predicted outcomes
35:38 Incorporating mathematics
45:27 The race to the bottom
56:18 Silicon Valley culture and problems with diversity
1:04:43 Differences in culture
1:06:56 Responsibility of AI research
1:09:22 Metric of fairness
1:11:06 Computing fairness
1:18:00 The complexity of these systems
1:19:43 Women in data science
1:21:48 Wrap-up