Superhuman Abilities: Can AI Empathize
AI will never feel, but it may learn empathy and adequate emotional response. Moreover, machines will soon understand and respond to our emotions better than our loved ones.
The power of machines lies in their superintelligence, the ability to perform the most complex computational operations on huge amounts of data in a split second. On the other hand, a person has much more modest capabilities and generally is an irrational being: our thinking is subject to distortions under the influence of emotions, external factors, and noise. For a long time, it was believed that despite the apparent weakness of humans in front of machines in terms of reason. Humans surpass machines in their ability to feel. The machine does not have a biological body, which means that it will never understand how the limbs grow cold from fear, from anger it takes a breath in the chest, and from love, “butterflies in the stomach” flutter. Consequently, AI will always lag behind humans in empathy and empathy. But is it? Also, emotion control is needed in sports betting. Try your capabilities on this 22bet.cm/en/prematch.
Modern technology shows that AI can understand much more about the nature of human feelings than we think. If trained properly, machines can more accurately recognize and respond to emotional states than humans.
Emotional thinking will allow machines to move to another level of interaction with people. They can become better friends and life partners than the people around them are. Moreover, the improvement will allow machines to make moral decisions since empathy underlies the definition of good and evil, good and bad deeds.
Also Read: Why is Java a Platform Independent Language?
How Machines Can Recognize Emotions
For centuries, all believed that human feelings are incomprehensible, not amenable to mathematical measurement. It is not for nothing that we use words such as “inner world,” “mysterious soul,” “abyss of subconsciousness” to describe feelings. Modern brain research is making it possible to demystify human feelings – and lay the groundwork for artificial intelligence in this area. Although machines cannot sense emotions, they will understand them.
According to neuroscientists, emotion is an ancient decision-making mechanism that developed in all mammals even before forming the cerebral cortex, responsible for conscious thinking. The emotional structures of the brain act as a primitive command center: in all mammals, from rats to humans, they contain a pleasure center and a fear center that elicit a fight-or-flight response. For example, the pleasure center is activated in the brain when we see food or a partner ready to mate – a pleasant sensation suggests that we get closer to this goal. At the sight of threatening objects, for example, snakes or a rapidly approaching bus, the center of fear is activated, which dictates that we jump aside in a matter of fractions of a second. Everything we experience as subtle vibrations of the soul’s strings is in these rather basic biological processes.
The biochemistry of emotions is universal among people of different nationalities, gender, and age. The “happiness” hormone dopamine makes all people feel a surge of strength and energy to move towards a goal; cortisol forces the neuromuscular system to mobilize to fend off a threat – to run faster or hit harder. An increase in heart rate increased pressure, and increased blood glucose levels are the first harbingers of an emotional reaction. And this is what today’s gadgets like fitness bracelets or Apple Watch can perfectly track.
The leading researcher in the field, University of California professor Paul Ekman, has observed that even expressions of emotion are universal across cultures. It’s not for nothing that we have universal international emojis. Anger is expressed by frowning and clenching the jaws to protect the eyes. And mouth in a collision. When disgusted, the nose wrinkles and the tongue protrudes slightly to prevent the poison from entering the mouth and nose. Fear is expressed in wide eyes to increase the volume of perceived information and speed up the response to the threat. Emotion recognition by facial expression has been developing as a director of machine learning for quite some time.
People have similar not only facial expressions but also verbal expressions of emotions. Despite the wide variety of terms used to refer to feelings, it is quite possible to train a machine to recognize emotional states from speech using big data.
The Machine Can Recognize Emotions More Accurately Than a Person
The human ability to recognize each other’s emotions is greatly exaggerated. We often have difficulty understanding the feelings of even those close to us. For example, we may notice that a partner avoids communication with us – he is silent, withdraws from the conversation, wants to be alone. What’s the matter? He may be angry. Perhaps he is upset or offended by something. He may be tired at work or lost in thought. We often guess and make assumptions because we cannot understand the real feelings of the other person.
Moreover, we are often unaware of our own emotions. We skip the signals of incipient emotional reactions and feel them only when emotions “overwhelm” over the edge, when we scream out of anger, cry out of sadness, or cannot find a place for ourselves because of solid anxiety. But emotion is a process that grows gradually and is always expressed through signals in the body.
Machines can be much more accurate in determining human emotions, mainly if you teach them to synthesize information from different sources – to compare facial images with speech data and metrics of physiological processes in the body.
You can deceive your boss with apparent calmness when he criticizes you, but the devices will determine by the pulse or trembling voice that you are overwhelmed by great excitement. When your partner offers to spend the weekend with their parents, you will say “How wonderful!”, But the clock will show that you are depressed.
Empathy for Machine Intelligence
In the classical sense, empathy implies empathy and empathy for another person. It seems incredible that a machine devoid of sense organs would be capable of this. However, seeing the capabilities of artificial intelligence, we can say that a machine can understand your feelings in its language of numbers even more than another person.
As soon as the machine recognizes the presence of a particular emotion by a set of signs, the second step is to develop an adequate response mechanism. Evidence-based psychotherapy in recent decades is precisely aimed at identifying the most effective universal techniques that could follow a specific algorithm.
Many devices already offer a basic set of emotion regulation – for example, taking a breath, thinking good things, or doing meditation. In the future, we will undoubtedly expect the development of conversational intelligence, which will go beyond simple advice towards full-fledged dialogues based on an analysis of the user’s situation. AI “empathy” will take the interaction between man and machine to a whole new level. Not only will the machine react more “naturally,” but it will also be able to make decisions that are in the realm of ethics and morality.
As neurological studies of MRI show, moral choice is always closely related to emotional empathy – the activation of emotional brain structures. Harming another person causes unpleasant feelings in living beings, while mutual help among community members generates positive emotions. Evolution has taught living beings that selfish actions to the detriment of community members (“immoral” actions) lead to mutual loss and extinction of the species, and actions that take into account the interests of fellow tribesmen (“moral actions”) contribute to gain and the continuation of life.
This finding is supported by research on the brains of psychopaths. Psychopaths are extremely rational people, but they have impaired emotional brain structures responsible for empathy – empathy. That is why they can kill, rape, deceive – and at the same time rationally consider and plan their atrocities. They are not touched by the experiences of other people whom they harm.
A machine that cannot empathize will act, in essence, like a superintelligent psychopath. Her actions will be subject to a super-rational logic that excludes consequences for other people from the calculation. This aspect is most frightening for humans when they think about the era of smart robots.
A machine that is programmed for empathy will make more humanistic decisions. A machine will assess the quality of the decision from the standpoint of practical benefits and the perspective of the influence on other people’s feelings, the degree of their suffering. With the appropriate command, the robot will execute the command to throw an object out of the window but not do the same with a person.
Risks and Dangers
People are already becoming attached to the simplest robotic devices if they give them positive emotions and help in everyday life. The American military buried the robot Boomer as a combat friend who found explosive devices and saved many lives with honors. Robot dogs have become very popular in Japan: their owners get together, buy clothes for them, and even maintain a separate cemetery for spoiling. Twisted devices. One wedding is already known – a Japanese man married a hologram robot. The owners of robots give them names, gender, take them for a walk, attribute character traits.
One can only imagine how much the attachment to the robot will increase, which will carry out the programmed commands and conduct conscious conversations with its owner. We have yet to realize the risks of interaction between a person and a program that recognizes and influences human emotions.
