A new kind of hacking is worrying the makers of autonomous vehicles and even the US military. These so-called ‘adversarial attacks’ can use information invisible to us to hopelessly confuse an AI entity. Their most dramatic demonstration was the recent placing of small stickers on the road to cause a Tesla to steer into the opposite lane of a two-way street. The implications are extremely worrisome both for self-driving carmakers and any military that thinks it can send autonomous drones into battle any time soon (though that might not be such a bad thing…). Read more here.
Today’s Simple AI – AI made easy