From the course: AI Accountability: Build Responsible and Transparent Systems
Join today to access over 24,800 courses taught by industry experts.
Discrimination in implementation
From the course: AI Accountability: Build Responsible and Transparent Systems
Discrimination in implementation
- [Presenter] It's common knowledge that AI has been associated with some shocking forms of discrimination and control, including racism and minority oppression. But it's not just a matter of bad people doing bad things, even though that is sometimes the case. There are other less obvious factors that can contribute to this problem. So for example, when people write their AI algorithms, it is a very frequent practice to use code packages that are developed already by other people. The problem is these code packages, which might, for instance, add facial recognition, you didn't develop those, and you may not know what goes into them. And consequently, there's a high potential for unchecked bias in the AI packages. Second, there can be the matter of ignoring AI errors from minorities. If you only look at overall accuracy and then you miss the high level of mis-categorization for people in numerically smaller groups, you're doing a disservice to your client base as a whole. And then…
Practice while you learn with exercise files
Download the files the instructor uses to teach the course. Follow along and learn by watching, listening and practicing.
Download courses and learn on the go
Watch courses on your mobile device without an internet connection. Download courses using your iOS or Android LinkedIn Learning app.