As AI technology proliferates, fears increase about how much we can trust it and the things they create. Despite fears (largely) being unfounded, researchers at the University of Chicago have found another way to abuse it, by training a 'neural network' to write reviews as well as humans can.