An organization dedicated to the safe development of artificial intelligence released a "breakthrough paper" it said will help humans better control the technology as it spreads. "We can’t trust AIs ...