AI in the News - #15

March 2, 2016 | Earl Wajenberg

They had better be trustworthy  (New Scientist)

... because we do, in fact, trust them.  In a psychological experiment, people were asked to work on some other task that was the ostensible purpose, but then the proceedings were interupted by a fire alarm and a bucket-like robot lit up with a sign labeled "Emergency Guide Robot" and proceeded to lead the people...

... away from a clearly marked exit.  And almost everyone followed.  Even if the robot "broke down" in the middle and had to be "repaired" by experimenters.  Even if the robot led the people toward a darkened room with a partially blocked door. 

Asimov's Second Law of Robotics is that "a robot must obey all orders given it by a human," and it appear to cut both ways.

Oops.  (New Scientist)

One of Google's self-driving cars has been in a minor traffic accident.  Google accepts responsibility.  No one way hurt.