How Much Will A.I. Surprise Us?

Why think about what neural networks (and AI in general) can do that we can already do, when he real question that we should be asking is this: What will A.I. be able to do that we can’t even dream of?

When we think about Artificial Intelligence, we often consider its potential in relation to the realms of our possibility – what will it be able to do that we can do? To my mind, that is entirely missing the point of an artificial “neural” network that is infinitely more powerful than the percentage of our brains that we are able to access at any one time. The real question that we should be asking is this:

What will A.I. be able to do that we can’t even dream of?


I’ll give the simplest example that I can. Way back in the early 1980s, there was a computer game called Breakout, where a horizontal paddle (bat) could be moved at the bottom of the screen, bouncing a ball up at tiles every time. When a ball hit a tile, it disappeared, and the ball rebounded back to the bottom. Repeat ad-nauseum until you have destroyed all the tiles. Some of you might know this game as Arkanoid. It was a game of simple geometry and fast reflexes, but there wasn’t too much strategy involved.

Then DeepMind unleashed their A.I. on it. When happened next stunned veteran gamers from all over the world. The A.I. found a different way to win – a way that in 30 years of passionate gameplay had never been discovered. The game couldn’t be simpler, but the A.I. had found the perfect way to win.

To me, this is a little shocking, but it is also incredibly exciting.

When A.I. is let loose on the simplest of problems, it comes up with something different. Just imagine what it will be able to do when we truly let it loose. How much is A.I. involved in solving the real issues of our planet – debilitating diseases, climate change and extremism to name but a few? Maybe one day A.I. will be curing us, healing our planet and helping us all get on that much better? At least, it will if we give it the chance to find the as yet invisible answers.

The key issue here is trust.

Do we really give A.I. the keys to the castle? What other conclusions could it draw? What if these conclusions put into question the future of the entire human race? How dangerous could it be if we allowed it deep enough down the rabbit hole? The answer is that we don’t know until we try, but with huge advances in the area, this day is not so far away. Perfecting a computer game is one thing, but, one day, machine learning could turn out to be the saviour of our planet.

Working in Big Data recruiting, that is an exciting prospect, but it may well be more real than we think. If we are open to A.I. surprising us, I am certain that the revelations will rock our world – hopefully for the better.

I only wish it would give us more insight into the pros and cons of a potential Brexit!

Original. Reposted with permission.