Responsible AI and Fairness Practices

#16 - Responsible AI and Building Trust in AI - Liu Feng-Yuan

Topic 7 of 16

Published Nov 23, 2020 · Clip 25:10–28:49 (3:40)

Feng-Yuan lays out Responsible AI practices: track what data, parameters, and assumptions go into training; apply software engineering hygiene like version control; use explainability methods to open the AI black box; enforce fairness by monitoring whether models err differently across demographics like gender or race. He warns that ML systems can silently learn wrong things while remaining 99.9% available, requiring different SLAs than traditional software.