Don't worry about AI becoming sentient. Do worry about it finding new ways to discriminate against people

Laura Edelson, a postdoctoral researcher in computer science at NYU Tandon, warns against using AI systems as a get-out for people who use them when they turn out to be discriminatory. "A common use case for machine learning systems is to make decisions that humans don't want to make as a way of abdicating responsibility. 'It's not me, it's the system'," said Edelson.