AI: Cheating
When I started working on artificial intelligence, experts debated whether or not, someday, a computer could play chess at a high level. Could a computer possibly be the world champion? Today, a computer has been chess champion for ages, and for $5 a month you can play as many games as you like against a computer that has a chess rating anywhere you please, from rank beginner to grandmaster.
When it plays chess, AI frequently seems inhuman. AIs make moves that people would be unlikely to find, both because they search more deeply than people, and because they don’t spend much time thinking about what they had been planning to do before you moved your rook. It’s often fairly easy for good chess players to discover cheaters — online players who are using a second computer to ask an AI for the best move.
Grandmaster Daniel Naroditsky, 29, recently died in circumstances that suggest suicide. He had been dogged for some time by accusations from former world champion Vladimir Kramnik that Naroditsky was cheating.
Everyone now assumes that naturally the computer is far more capable than the very best chess players.