Until recently, scientists argued that even the most perfect computer will never have the ability to think and reason on its own, because, by and large, it is just a set of microcircuits that allow the device to perform complex, but standard operations, put into it by people.
Nobody, of course, will put a program for the destruction of humanity into a computer. Although purely theoretically … anything is possible.
So far, no country in the world has technologies that could function without human control. Nevertheless, similar developments are underway. So, experts from the Georgia Institute of Technology have developed a special program that will teach robots … to lie. True, only to each other.
“Similar tactics can be used for robots guarding military depots or ammunition on the battlefields,” the authors of the project report. “This will help them deceive the enemy and buy time before reinforcements arrive.
The well-known company Google uses in its software algorithms of the so-called deep learning, which simulate the activity of neurons in the cerebral cortex. This allows programs to recognize certain patterns in sounds, images, and other virtual data.
A team of researchers led by University of California chemistry professor James Giemsewski, with the support of Defense Advanced Research Projects Agency (DARPA), has developed technology that allows the creation of fully autonomous robots capable of "thinking" like a human. They will work not on the microcircuits we are used to, but on nanowires that form billions of connections, similar to the neurons of the human brain.
This will help them remember and analyze the information received in the same way as people. These robots will not need human control. For example, aircraft equipped with these systems will study the terrain and plot routes without the participation of a living pilot.
The head of Palm Computing Jeff Hawkins, author of the book "On Intelligence" (2004), in turn, develops a machine learning system that predicts patterns of energy consumption and the likelihood of failure of various devices. However, he believes that machine intelligence will never be able to evolve to the human level, since even deep learning will not allow it, say, to experience emotions.
Promotional video:
Suppose the system can easily recognize an object such as a cat, but does not understand when this animal is doing something funny, and when it just sits in place … But perhaps the main danger of the impact of artificial intelligence on people is precisely the lack of humanity …
Campaign against robots
Recently, the global campaign "Stop the killer robots" is gaining momentum. Its supporters are convinced that the development of artificial intelligence threatens humanity with destruction. The campaign is led by Jody Williams, founder of the International Anti-Personnel Mine Ban Movement, who won the 1997 Nobel Peace Prize.
Experts say that in just 20-30 years, armies will not consist of humans, but highly organized robots. New technologies will make military conflicts practically bloodless; in any case, they will help to reduce the number of human casualties to a minimum. However, Jody Williams believes that the production of such "death machines" is a violation of the basic principles of humanism and human rights.
“I don’t want to be destroyed by a programmed machine,” she says.
Do you think this is nothing more than unfounded hysteria? Take the very recent case of the da Vinci surgical robot, the brainchild of Intuitive Surgical. This model is now installed in clinics around the world. It is believed that the invention greatly facilitated the work of surgeons, who sometimes have to perform complex operations for many hours without leaving the operating table …
Da Vinci is equipped with four mechanical arms, a set of cutting tools and a mini-telescope. The robot is controlled by live doctors using a special console, focusing on the three-dimensional image of the operated organ displayed on the monitor.
The robot is programmed to make miniature incisions, which reduces blood loss, reduces the risk of bleeding, and also shortens the overall recovery period after surgery. The device is actively used in such operations as organ transplantation, mitral valve repair, gastric bypass surgery, removal of the gallbladder and prostate gland, and many others.
At first, the doctors were delighted with the cybernetic "colleague", which really simplified their hard work in many ways. But recently, a number of incidents involving the use of "da Vinci" made doctors on their guard. So, two people died under the knife of the robot: one - as a result of an accidental cut by the robot of blood vessels, the other - after surgery on the spleen. Colon perforation during prostate surgery and trapping of internal tissues during rectal surgery were also recorded.
“We had to completely turn off the system in order to release the grip of the apparatus,” says the doctor overseeing the operations.
In addition, during a hysterectomy (removal of the uterus) operation, the robot struck the patient with one of its mechanical limbs. As a result, the surgeon had to refuse the "services" of the apparatus and operate on his own.
“Research results show that the risks associated with using a robot are significantly underestimated,” said surgeon Martin Macarius from Johns Hopkins University. - The rapid introduction of the da Vinci apparatus was made without assessing its capabilities. We are aware of cases of catastrophic complications that have arisen after operations carried out with the help of mechanical hands.
Almost alive
Nevertheless, robots are already so firmly established in our lives that we often perceive them as if they were living beings.
Foreign researchers tried to determine how interaction with robots affects people on an emotional and neurological level. Thus, the volunteers were shown two videos. In one of them, a small robot dinosaur was hugged and tickled, in another they mocked him - they beat him and threw him on the floor. Then the experts evaluated the level of physiological arousal of the subjects after viewing, recording the conductivity of their skin. After all, when we experience strong emotions, we sweat more and our skin conducts electricity better.
The participants in the experiment noted negative emotions received as a result of watching a video with a robot, over which the violence was committed. At the same time, the level of conductivity of their skin increased significantly, indicating the state of stress experienced.
In another study, participants' brains were scanned using functional magnetic resonance imaging (MPT). This time they were shown videos, in one of which a person was the object of violence, and in the other - a robot. For example, in one video a man tried to strangle a woman with a plastic bag, in another they did the same with a robot.
When volunteers were asked to observe a human and a robot that were being treated kindly, the scans revealed similar neural activity. If the object was mocked, then the subjects were more sympathetic to the person, and not to the robot. “We believe that robot abuse causes almost the same emotional response as human bullying,” said study leader Astrid Rosenthal von der Puten from the University of Duisburg-Essen (Germany). - However, we still have more sympathy for the person.
“The fact that a person shows sympathy for a robot is not surprising, because robots look and behave like humans or animals,” says robotics engineer Alexander Reben.
The researchers note that showing empathy for robots occurs in a wide variety of situations and contexts. For example, soldiers on the battlefield demonstrate an emotional connection with robots. True, the more realistic artificial creatures seem, the greater the degree of sympathy for them can be observed. It doesn't work, though, if robots are too human-like. We are instinctively rejected by our artificial "copies" …
Ida SHAKHOVSKAYA