Is AI discriminating?
Is AI discriminating?
We are in an era where the future of humankind is intimately linked with the evolution of the Internet. The latter is in constant change because of the new technologies that are associated with it and that keep appearing.
In my application for the Mozilla Fellowship I mentioned yesterday, I wanted to investigate how AI and more specifically the algorithms and processes leading to its implementation can lead to various types of discrimination, whether based on race, gender, sexual orientation, religious affiliation or other types of criteria.
The main issue I want to addressed is the disparity in the representation of social groups or minorities (women, people of color, LGBTs, etc.) in the data sets and automatic learning processes that precede the implementation of an AI
For me, Internet is one of the best things that has ever happened to humanity. And I believe that it must, as well as all the applications, technologies, services or products that emerge thanks to it, be of benefit to all without distinction. My application to Mozilla Fellowship was therefore for me a way to realize, at least in part, this vision.
Here are some resources that I found useful while working on the application:
Amazon : l’intelligence artificielle qui n’aimait pas les femmes
La discrimination, côté obscur de l’intelligence artificielle
Artificial Intelligence’s White Guy Problem
Handling Discriminatory Biases in Data for Machine Learning
AI programs exhibit racial and gender biases, research reveals