Humans can trick Google’s self-driving car

For people still leery about the booming field of artificial intelligence, the good news about last month’s traffic accident involving a Google-made self-driving car is this: Robots still haven’t figured out us human beings. That may be of little comfort if you’re in the market for an autonomous vehicle. The Feb. 14 accident in Mountain View happened when a Google car tried to enter a lane under the assumption an oncoming bus would yield. A misunderstanding ensued, as did a very minor accident, Google’s first. The much deeper problem exposed by the collision, according to an article at Slate, is that humans will be able to trick and otherwise take advantage of robot cars. Why? Because they (still) can’t divine our intentions. “Even the best AIs are easy to fool,”…


Link to Full Article: Humans can trick Google’s self-driving car