Machine Learning is defined to be ‘super-human’, but how do we deal with mistakes and errors when the stakes become higher? Using a neural network model to detect fraud has tremendous benefit in keeping fraud under control and the cost of a false positive (blocked legitimate transaction) is low. But as we start looking to machine learning to drive our cars, perform surgery, or monitor our movements, how will we become comfortable with higher-cost mistakes?
At FICO World 2018, I sat down with Chess Grandmaster Garry Kasparov to discuss the hot topics in the world of AI, and asked: Should we expect AI to make perfect decisions?
Do I agree with Garry? Read my own thoughts on AI and related analytic topics on Twitter @ScottZoldi