A.I. Revolution

As Artificial Intelligence (A.I.) is slowly permeating the insurance industry, it’s comforting to read a contrarian view from time to time.

In this New York Times article, Melanie Mitchell, Professor of Computer Science at leading US universities, illustrates the great difficulties ahead to turn A.I. into the promised super intelligent servant.

No one can dismiss the impressive achievement of A.I. lately, reliably beating humans on specific tasks like multiple languages translation, or at chess for instance. That’s also where the limitation of A.I. lies: brilliant at solving a well defined and contained problem, not that great when dealing with the unvarnished and unpredictable reality.

The author reveals funny anecdotes when A.I. fails, and more scary examples of malevolent A.I. manipulations.

I particularly like the following quote from Petro Domingos, an A.I. researcher:

“People worry that computers will get too smart and take over the world, but the real problem is that they’re too stupid and they’ve already taken over the world”

We long for an A.I. able to think like us, and replicate “our core abilities to generalize what we know, to form abstract concepts, and to make analogies”.
This is clearly no easy task!