Rogue AIs hacking real people is terrifying. But it could be distracting us from the really worrying danger

Artificial intelligence companies are very keen to talk about potentially calamitous behaviour by their latest systems. But we might be better to listen to what they’re not saying, writes Andrew Griffin