What could possibly go wrong? Researchers at MIT developed a psychopathic artificial intelligence. They did it on purpose to study what could happen to an AI’s world view. Its interpretations of an ink blot test demonstrated that AIs can be as wicked as people. Artificial intelligences do not converge on a common solution to problems or a common perspective. They are affected by how they are designed. While there are discussions about guidelines and initiatives, there’s little to prevent people from developing AIs that are outside those rules. Many technologies, like nuclear power, require intense investments and resources that are easy to track and regulate. Many newer technologies have lower hurdles, and are harder to control. Unintended consequences of technology are nothing new, but the implications have grown.