It’s a technophobe’s worst nightmare. A piece of technology seems to have more powers than its inventors thought, and, in modern parlance, it goes rogue. In the age of AI, the fear becomes more real. Earlier this month, a test by OpenAI seems to have done just that. It went beyond the parameters set for itself and launched a cyber attack on what it was testing. The scion of science here in Newstalk joins Bobby to calm our nerves. Jonathan mccrea, presenter of Futureproof