Researchers in the U.S. military are working to combat what they call "adversarial artificial intelligence." That's when someone hacks into an AI system to transmit the wrong information.

• 
•