Ahead of the U.S. election in November, social media has been on high alert for a technology that threatens to disrupt fair voting. As Mark Niu reports, AI and deepfakes are getting more sophisticated, but researchers are working on new ways to detect them.