In his first account of his landmark chess defeat in 1997 against IBM supercomputer Deep Blue, Kasparov explains why, 20 years on, artificial intelligence is good for humans.