Tumblelog by Soup.io
Newer posts are loading.
You are at the newest post.
Click here to check if anything new just came in.

December 04 2017


Battling algorithmic bias

Part of the problem is that the complex mathematical models behind today’s algorithms aren’t all held to the same ethical standard. It happens that a large proportion of software development and oversight is done by men. Though perhaps unintentional, the potential for products to inadvertently adopt some bias or emotion increases.

The algorithms behind online lenders, for instance, exhibit much of the same gender biases as traditional lenders. These new lending platforms are pulling in all kinds of alternative data such as purchase and browsing history, on top of conventional requirements of income, education, and residence, to determine a borrower’s creditworthiness. Many of these determinants regularly favor men despite a growing body of research that proves women tend to save more and make more on-time loan payments.

As a result, women entrepreneurs still receive fewer and smaller business loans with higher interests rates compared to men. This disparity in financing denies women an opportunity to make an equal contribution to economic growth.


Don't be the product, buy the product!