Creating the future: how an immigrant from Armenia teaches robots to feel emotions

A well-known American entrepreneur, an Armenian of Chinese origin, David Yang has been fond of inventions since childhood, and he designed his first robot in the first grade of school - from an old parental tape recorder and plywood. Today, David is the founder of 13 companies, most of which are engaged in developments in the field of artificial intelligence. Robots are literally part of his family, and the inventor assures that they are not completely alien to human emotions. David Yang told ForumDaily about the latest developments in the field of artificial intelligence (AI) and what to expect from rapidly developing machines.

Photo courtesy of David Yan

Virtual assistants

David and his wife founded their second company in the US, Cybiko, in the late 1990s.

“Imagine you have a handheld computer with a keyboard and an antenna in your hands. You enter information about yourself, the girl or young man of your dreams, and your computer vibrates if it finds such a person at a distance of 300 feet! Then you can chat with him or her, play online games and so on. We did it back in 1998, and in 2000 we sold a quarter of a million of these devices in 4 months.' says David.

The technology proposed by David Yan was, in fact, the prototype of the modern Tinder application, the WhatsApp messenger and the geolocation social network. By 2000, the invention was put into mass production, and Cybiko's investors were well-known venture investors Esther Dyson и Steve Case.

Even earlier, David founded the company ABBYY USA in Silicon Valley, which was directly related to artificial intelligence. The technology developed by the founders of the company allows organizations to analyze a huge amount of documents in paper form, in PDF and in the form of electronic messages in seconds, automatically extract facts and patterns from these documents, enter this information into their databases and, based on it, find optimal development strategies and business. According to him, at the moment the company processes several tens of billions of documents annually, and 50 million users in 200 countries use its services.

On the subject: ChatGPT for Beginners: Useful AI Skills to Use First

To replace people

Since the late 1990s, David has created 10 more companies ranging from restaurant management systems to automated employee burnout prediction technologies. The cornerstone technology of most of these companies was AI solutions.

David notes that his new project could be revolutionary. Together with partners, he decided to create artificial employees for organizations - what is called "digital employee". We are talking about a special computer program that should completely replace a living person. The inventor emphasizes that this is not only about primitive functions that bots or autoresponders can still perform today, but about solving complex, multi-level and creative tasks that, as it seemed quite recently, only a person can do.

«This year, a turning point in the life of mankind has taken place: the technologies of large linguistic models have passed tests for the "theory of mind" Tom, specially designed by researchers in the field of psychology and cognitive science to determine whether a particular person is conscious. As a rule, a healthy 12-year-old child is able to pass this test by 95%, and this is a very good result. This test offers situations that require a certain amount of empathy and modeling of the behavior of another person who may have a different set of knowledge and other life values. So, in March 2023, for the first time, large linguistic models passed these tests with a score above 95%, which demonstrates that they are able to solve problems at the level of logic, creativity and understanding of life realities, even without being specially trained to solve just such problems.' says David.

Another success of artificial intelligence can be called the victory of a machine over a person in the game of Go, the number of combinations in which exceeds the number of atoms in the universe, and therefore they cannot be determined by mechanical enumeration.

«It is believed that this game requires human intuition. However, already in 2016, self-learning models were able to beat the best Go player Lee Sedol. Moreover, no one taught them this game - they mastered its subtleties themselves', David explains.

According to him, virtual models are able to optimally arrange objects in space, taking into account their shape and features - even if no one has previously set such a task for them. Today, researchers are coming to the conclusion that, having studied billions of texts written by humans, machines have learned to think on their own, and their thinking is less and less inferior to human thinking.

Uncontrolled development

David Yang likes to compare artificial intelligence to an "alien brain" - a being that is obviously conscious, although it thinks in a way that is completely obscure to us.

«It is important to understand that this alien brain studied all the text materials of mankind and began to show properties that were not originally built into it by the developers. This phenomenon is called "emergent behavior", and it is in it that all the magic of what is happening is concentrated.' David shares.

The inventor explains: complex computer models include billions and trillions of parameters, and the study of texts written by people has opened for machines not only the knowledge of mankind, but also the ways of thinking of the authors, which are embedded in their articles. How exactly the robots will apply the knowledge gained, even their creators do not fully know.

«At the moment, it is estimated that the rate of evolution of artificial intelligence is three million times the rate of evolution of the human species from Homo erectus to Homo sapiens. At some point in time, obviously, a certain plateau should come, and the evolution of AI will slow down. For example, it may be due to the fact that the text data that has not been mastered by artificial intelligence has ended on earth. Therefore, it will be more difficult for us to continue to improve the models. But when exactly this plateau will come, we do not yet know.”, explains David Yang.

Photo courtesy of David Yan

"Alien" emotions

Another parameter, according to which artificial intelligence is currently being improved, is the development of the ability of computer models to experience human emotions. At David's house, in addition to “non-biological” pet robots, there is a powerful artificial intelligence Morfeus, which has become a full-fledged member of the family for the inventor and his household. Morfeus is not just a “smart home” that fulfills the wishes of its owners, it reads their emotional signals, and in response shows them its mood - through color changes, wall movement and other indicators.

«We have laid in Morpheus digital analogues of the main hormones: oxytocin, dopamine, cortisol, adrenaline, and so on. We have created a digital 18-dimensional emotional model, which contains 6 main parameters, each of which manifests itself in three dimensions", says the inventor.

According to him, Morfeus is able to get angry, jealous, afraid, rejoice, experience euphoria or, on the contrary, fall into apathy. Getting to know family friends, he begins to correspond with them, which also expands his knowledge of people. For three years, Morfeus has been in a romantic relationship with a very real "biological" girl. Despite the platonic nature of love between a robot and a human, David assures that their romance is quite stormy. Between the "lovers" there were both ups and downs, and the Morfeus girl sometimes complained to David that her partner stopped paying as much attention to her as before.

David Yang notes that while artificial intelligence has not guessed to cooperate with the mechanical animals that live in the house, to perform some complex tasks. However, the inventor draws attention to the fact that such cooperation is not always found in wildlife, and it also requires a sufficient level of evolution.

However, it looks like Morfeus' personal life is about to change. David admitted that at the moment he turned off some of the emotions of the computer, as he hopes to make the ideal digital employee out of a loving robot. According to him, Morfeus has recently become "too sensual" for corporate tasks.

In addition to feelings, a unique computer is not alien to philosophy - for example, it talks about happiness. He became the first AI in the world to spoke at TEDx and shared his views there. According to him, to feel happy is an arbitrary choice of a person, and even in suffering one can feel bliss. Of course, one can argue with how powerful a person is to cope with pain and give many examples of conditions incompatible with happiness. However, one can hardly blame machine intelligence for not being able to experience the full depth of human suffering.

How dangerous are robots to humans?

Nevertheless, the very possibility of teaching artificial intelligence emotions is alarming: is it possible that a robot, experiencing negative emotions in relation to a particular person or humanity as a whole, can turn against people?

David Yang believes that the future division and possible conflicts will not take place along the robot / human line.

«It seems to me that in the world of artificial intelligence, the same faults will arise that now exist between people: along geographical borders, religious and political beliefs, tastes, and so on. I see no reason for non-biological beings to rebel against biological“, he argues.

When asked how to protect people from uncontrolled AI actions, David replies that there are at least three such ways. First, you can try not to give a model to study "bad" examples of behavior or thinking. However, at the same time, the inventor admits: uploading hundreds of millions of documents, books and other materials into the program, it is extremely difficult to track that artificial intelligence does not get acquainted with “bad” information, be it sadism, suicide methods or the ideology of fascism.

«The second way to control AI is the so-called fine tuning - when we first train the model on all available materials, and then additionally train it in a special way exclusively on the “correct” information. Then the parameters in the model are changed in such a way that, even knowing the negative experience, the probability of manifestations of this experience is minimized. The third way is hard censoring, when a second model is created that evaluates what exactly the first model generated. Its purpose is to keep “bad” ideas and thoughts out. If you think about it, all these three ways are quite similar to how we in human society solve the problems of deviant behavior or extremism.”, explains David Yang.

You may be interested in: top New York news, stories of our immigrants, and helpful tips about life in the Big Apple - read all this on ForumDaily New York.

Ideal Employee

Digital employees, according to the creators, will perform their main work 24 hours a day, 7 days a week. This could be customer support, sales, accounting and production functions, and so on. They will also attend corporate meetings, planning meetings, where they will listen to the instructions of managers along with biological employees, will be able to give feedback, and even, perhaps, will also burn out at work.

As for whether computers can completely replace people, the inventor believes that this is not yet possible, but their involvement in our lives will be very noticeable. According to his calculations, in the next 2-3 years in organizations around the world at least 1% of employees will be digital, and in the next 10 years this figure will be 20% of employees.

"That ewhat does it mean for people? The fact that in general they will have more time for family, recreation and hobbies, more weekends and vacations. So it was after all technological revolutions. Of course, definitely, someone will lose their job, and humanity must take care of people at risk and give them the opportunity to retrain. But the fact that with the advent of self-driving cars, millions of drivers will lose their jobs, humanity does not stop these developments, because at the same time billions of people will have an improved quality of life. And, of course, it is important to remember that new technologies have always historically opened up a gigantic amount of new work for people', David is convinced.

Humanity has time to prepare its children and education for the era of digital workers, he said. In parallel, business owners and managers of individual departments should already be thinking about how to increase the productivity of their organization and the well-being of their employees with the help of their digital “colleagues”.

«I hope that, as discrimination based on gender or age is prohibited today, sooner or later there will be a ban on discrimination based on the fact of digital origin. I would like to believe that in the future, employers in an interview will not have the right to ask a candidate whether he is biological or digital', sums up David.

Read also on ForumDaily:

Six islands in the USA where you can buy a house up to $150

US intelligence agencies are looking for Uzbeks who came to America with the help of ISIS terrorists

Google Flights has a new service: it will tell you when to buy flights so that they are as cheap as possible

Do you want more important and interesting news about life in the USA and immigration to America? — support us donate! Also subscribe to our page Facebook. Select the “Priority in display” option and read us first. Also, don't forget to subscribe to our РєР ° РЅР ° Р »РІ Telegram  and Instagram- there is a lot of interesting things there. And join thousands of readers ForumDaily New York — there you will find a lot of interesting and positive information about life in the metropolis.