Google teaches AI to fool humans so it can learn from our mistakes

Google’s AI researchers are trying to figure out the best ways to fool both humans and computers in order to develop defenses against adversarial attacks on image recognition systems.

Sourced through Scoop.it from: thenextweb.com