Never play games with a bot(robot)—it will find a way to cheat if it can. A team from OpenAI, an artificial intelligence lab in San Francisco, has developed artificially intelligent bots that taught themselves to cooperate by playing hide-and-seek. The bots also learned how to use basic tools to help them win.
Bowen Baker at Open AI and his colleagues wanted to see if the team-based dynamics of the OpenAI Five could be used to produce skills that could one day be useful to humans.
The researchers set their bots loose in a simulated (模拟的) environment filled with fixed walls, movable boxes and ladders, and left them to play team games of hide-and-seek. The bots each had their own view of the world and couldn't communicate with each other directly.
At first, the hiders simply ran away. But they soon worked out that the quickest way to stop the seekers was to find objects in the environment to hide themselves from view. The seekers learned that they could move boxes around and use them to climb over walls. The bots then discovered that being a team-player -passing objects to each other or cooperating on a hideout -was the quickest way to win.
But the real surprise came when the bots started making use of problems or faults. Seekers found that if they pushed a ladder towards a wall, they could launch themselves into the air and spot hiders from above. Hiders found that they could get rid of the ladders by pushing them aside.
It shows that AIs are able to find solutions that humans miss, says Baker. "Maybe they'll even be able to solve problems that humans don't yet know how to." However, it is a large leap (跳跃) from virtual hide and seek to real problem-solving. "The main limitation is that it is in simulation," says Chelsea Finn at Stanford University.
