If spammers are able to write a program to read distorted text (and images) in CAPTCHAs (scrambled letters that attempt to block spammers), they have solved an AI problem,” says their creator Luis von Ahn of Carnegie Mellon University.