Fighting Bias in AI Starts with the Data

Author: Joe McKendrick

Publisher: ZDNet

Publication Year: 2022

Summary: The following article discusses the importance of data accuracy when building non-biased artificial intelligence (AI) models. At the heart of AI models is data, and so it is imperative to use high-quality, inclusive datasets that are not biased or skewed to deliver what the author refers to as “responsible AI.” I think this article is especially relevant to us as analytics students. It reiterates many of the points we have explored so far in our data ethics class discussions, and talks about how there is a lot more work to be done when it comes to reducing bias in the AI modeling world. While many of the IT executives in the survey referenced by the article agreed that data accuracy is a critical part of AI modeling, only a small number (6%) reported having achieved full data accuracy. This makes it all the more important to us, as analytics students, to utilize and advocate for the use of unbiased data as we move forward in our careers, regardless of which industry we work in.