On May 11, 1997, IBM’s Deep Blue computer stunned the world by becoming the first machine to beat Garry Kasparov, the reigning world chess champion and, probably, the greatest chess player of all time. Losing by one game in a six-game match (final score: 3½ – 2½) might not seem much, but it was seen as a major milestone in the long march towards “artificial” intelligence (AI). Even if the triumph of Deep Blue was really a victory of brute computing power and clever programming, a lot of people, including Kasparov, were shocked and wondered what really happened.
Since Alan Turing envisioned AI, technical progress has been measured by the ability to defeat humans in zero-sum encounters. In the last twenty years we have seen many examples of computers beating humans (in Chess, but also in Poker or in Go). However, in my opinion, this is not surprising or even scaring. We have known for centuries that machines are much better at some specific tasks: machines have better memory, calculate faster and make less mistakes.
Twenty-one years later, artificially intelligent systems of all kinds (driverless cars, robotic medical assistants, autonomous trading algorithms and room service robots) are becoming “natural” companion of humans. Think for instance at Siri (Apple), Alexa (Amazon) and Andi (Skype). AIs are providing higher-level functions and they are becoming full-fledged work team members.
Even more interestingly, AIs can now be equipped with an artificial set of “social skills” that allows machines to collaborate with both other kinds of machines, as well as humans. This means that AIs are learning to cooperate. And this is fascinating for me. Cooperation does not require enormous computational power, but a set of peculiar humans qualities: intuition, adherence to cultural norms, emotions, understanding social cues and signals from the environment and the situation. Recent research has shown that AI can be programmed with algorithm specifically designed to foster cooperation when AI is interacting with humans. The next step is to test if AI can sustain cooperation in large network of players and if humans can exploit machines when they interact.
*** Thanks for joining me today. I’d love to hear your thoughts on this topic. Please you “leave a comment” option below the title of this post. Finally, I would love it if you could use those little share buttons below to share my post with your friends.