We’re still a long way from real “artificial intelligence” – developing computers that can replicate the cognitive processes of the human brain. Yet, argues Colin Allen, with robots ever more able to complete certain tasks without real-time human input, it’s time to start considering how to program them not to act immorally.