Garry Kasparov gives his first public account of his landmark 1997 chess match with the IBM supercomputer Deep Blue, and explains why, twenty years later, he's become convinced that artificial intelligence is good for humans.