Facebook’s ‘Red Team’ Hacks Its Own AI Programs

Attackers increasingly try to confuse and bypass machine-learning systems. So the companies that deploy them are getting creative.

Read the original post by Tom Simonite