DARPA has actively started solving the problem and training robots to better recognize human gestures and “tricks”.
Extremetech, military veteran and former Pentagon analyst Paul Sharre is preparing to publish his book, but a funny story that happened to the DARPA-developed robot and the infantrymen who tested it has already been learned, writes Extremetech.
The Pentagon Defense Advanced Research Projects Agency (DARPA) is proud of its robot, for which scientists have invested a lot of money and effort. But when it came to testing the business model, fooling an AI robot proved to be child’s play.
The Marines were tasked with silently approaching the robot, which had previously been trained by experts and engineers to recognize humans and their movements throughout the week. But the soldiers easily coped with the task, deceiving the DARPA robot. One of them simply put a large cardboard box and calmly approached the robot, who could not recognize the “square” person. The other warriors robot somersault rolled over and he didn’t understand anything because the soldiers acted irrationally, not the way humans usually do.
As a result, the engineers realized that human cunning knows no bounds, and now they teach the robot all possible movement options and methods of camouflage using objects.
And experts note once again that no matter how much we admire artificial intelligence, it cannot do anything without the data we provide it.
Previously Focus He talked about how AI and robots will affect gunfights.
Source: Focus
Alfred Hart is an accomplished journalist known for his expert analysis and commentary on global affairs. He currently works as a writer at 24 news breaker, where he provides readers with in-depth coverage of the most pressing issues affecting the world today. With a keen insight and a deep understanding of international politics and economics, Alfred’s writing is a must-read for anyone seeking a deeper understanding of the world we live in.