Not all AI devices are created equal.
Some need only a small amount of data to perform useful tasks – others need to know everything about us. So, what makes people comfortable with one AI use case and uncomfortable with another? And how flexible is the line between them?