First it was chess; now humanity has lost the game of Go to the mighty computer. Thanks to breakthroughs in deep learning, artificial intelligence is on a rocket to the top. Yet scientists warn us against setting our sights too high.
It came as a terrible shock to the South Korean government. In mid-March, Korean Go master Lee Sedol was beaten 4-1 by Alphago, Google’s AI program. Two days later, Lee’s government responded by raising its budget for research into artificial intelligence to a trillion SK won over a period of five years, some 765 million euros and more than 50 per cent higher than originally budgeted.
But South Korea wasn’t the only one surprised by Alphago’s victory. Go was generally viewed as the last major hurdle in artificial intelligence. When IBM’s Deep Blue beat Garry Kasparov at chess in 1997, people could still point to the ever-growing advances in computational power that enabled machines to brute-force their way to game insight. But the number of possible positions on a typical Go board is so many orders of magnitude higher (in fact, it’s a googol times more complex) that, nearly twenty years later, it still can’t be managed.
This article is exclusively available to premium members of Bits&Chips. Already a premium member? Please log in. Not yet a premium member? Become one for just €20 for the first year and enjoy all the benefits.
