By us, buy us…Houston, we have biases!
Artificial intelligence is slowly but systematically invading our lives. Hardly a day goes by without an article or video extolling the capabilities of these new tools, which are supposed to solve all our problems, or at least make our lives easier and more efficient…
But every rose has its thorns, and artificial intelligence has its own. I’m talking about the systematic and sometimes dramatic biases that accompany its analyses, and their subsequent decisions which sometimes make the headlines.
Where do these biases come from? Why are they so difficult to anticipate? How can we try to detect them, and how can the tester help? that’s the subject of this presentation.