Artifakal Intelligence #2: “Machines are objective” (2/2)
Are AIs really that fair? (spoiler: no)
The introductory article of this project is available here.
This week, it’s Part II of the second article in the Artifakal Intelligence series. I’ll keep discussing the idea that: “machines are logical, hence they are objective”. Today, we’ll focus more on examples and possible solutions.
Last week, we saw why the concept of “objectivity” is sometimes complex to define. I also gave some examples of issues that can arise from the datasets and how these can determine unexpected and unwanted behaviors for the AIs. In this second part, we’re going to take a look at some real-life applications of AI where biases had a direct impact.
Note: if you haven’t read it yet, the first part of this article is here.
Why biased machines have an impact on our lives
Nowadays, lots of people are aware that robots and AI are everywhere — even when it is not obvious to the eye. What we can then wonder about is: how do the aforementioned issues transfer to our usage of AIs in the real world? Can these theoretical biases be harmful in practice?
