A New Kind of AI – TechTalks


Why are we about to create a technology that will combine the computer with the human nervous system into a single complex? How will a person change after connecting to an electronic machine? Can a computer system handle the flood of data from billions of living neurons? I will try to answer these questions in this article.

In the previous article “Individual artificial intelligence: a new technology that will change our world”, we talked about the fact that a new type of artificial intelligence will become a bioelectronic hybrid in which a living human brain and a computer will work together. Thus, a new type of AI will emerge: individual artificial intelligence.

By combining fundamentally different components (a living human brain and an electronic system), we will see how the emergence effect works (the birth of new properties that are not inherent in the separately combined elements).

Individual artificial intelligence as a bioelectronic hybrid will have fundamental advantages over both the biological brain and any computing system. We also talked about some technical characteristics of a new type of brain-computer interface.

The previous article raised several ethical questions and even concerns about our future. I will try to answer some of them.

What will happen to all of us? Who will we become in the end?

I will say what I think: we ourselves are going to change a lot. I don’t want to be moralistic and try to please everyone. As a doctor and scientist, I know that the development of science and the birth of new technologies cannot be stopped or delayed for moral or spiritual-philosophical reasons. A person with individual artificial intelligence will not become meaner or nicer, they will become stronger. I say this because I understand that the main source of human strength is not in the muscles, but in our head.

But the main question is how much stronger a person with individual artificial intelligence will be than us ordinary people. Apparently, hardly anyone in the modern world understands what this is about and does not realize the magnitude of the upheavals to come.

I am currently preparing an article in which I want to talk about the military aspect of this technology. I will not specifically analyze the perspectives of a fighter pilot, or a tank commander equipped with his own artificial intelligence. I mean a simple infantryman who, in symbiosis with the AI, will turn into something much more dangerous than a tank or a combat aircraft.

We await the inevitable transformation… which is already underway

If you look closely at the human essence, it becomes clear that the main difference between our biological species is the ability to change material objects and the environment around us.

Everything we create, houses, mechanisms, clothes, all this is the product of our intellectual efforts, the direct result of the work of our mind which, each time inventing something new, changes us a little.

The truth is that our inventions change us…

A clothed man is different from a naked man, and one who lives in a modern house is not like one who survives in the wild jungle. A person who drives a car after a while begins to feel different from someone who has walked all their life.

It’s the same with a computer

Look around you. We all spend a lot of time with computers, smartphones and various electronic devices. We are getting used to being in constant contact with the technical devices we create, and this started relatively recently. We are literally changing before our eyes, and the fact that at some point our brain will fully integrate with the electronic system is a fatality that will be the beginning of a new type of AI: artificial intelligence composed of a computer and a human brain.

Look around you and you will see millions of people with their attention riveted on the screen of a smartphone or laptop. All testify to the inevitability of our association with an electronic intellectual system.

I will say more:The moment we consciously took a stick and turned it into a tool, the creation of individual artificial intelligence became a matter of time.

Do you think personal AI is an unusual idea?

Indeed, for 65 years, we have been doing everything we can to make this concept a reality as quickly as possible!

Let’s remember together how it was

The very concept of artificial intelligence appeared in 1956 after the “Dartmouth Seminar”, a conference on artificial intelligence at Dartmouth College.

The original idea was to create a mechanism capable of replacing a human calculator. As part of this concept, he was supposed to create a machine capable of simulating the multiplicity of neural connections using an artificial technical neural network.

At the end of the eighties, instead of creating a rigid computer mechanism, information processing becomes the main objective. The main difference between the new systems was the use of flexible algorithms.

At the beginning of the 21st century, based on the growth of computing power, another new idea emerged, which we now call machine learning. The new algorithm not only retained its former flexibility, but also became capable of shifting the boundary parameters, thereby changing the very range of a possible response.

Soon, through the use of advanced computer systems, it became possible to machine learn and accumulate experience using huge databases and multi-layered technical neural networks.

This resource-intensive process has become known as deep learning. A good example of this relatively new technology is Google’s instant translation system. Instead of unraveling and clarifying logical connections in natural language structure, Google’s new model simply benchmarks ready-made versions of the original language.

The new method works on the principle of a black box and at the same time impresses with its precision and efficiency.

A similar technique was used by DeepMind when creating an AI to play the game Go. This system (deep reinforcement learning) uses two deep neural networks at once (one analyzes position, and the second analyzes the strategies used in the protocols of past games).

Please note that with each new step, we are moving further and further away from the idea of ​​creating an alternative for ourselves: technical AI. Instead, we’re trying to build a smart tool to help us solve specific problems.

The history of the past decades suggests that what we really need is not a competitor, but a tool – a new mechanism that allows us to solve various problems!

The new brain-computer interface will become such a tool, it is what will allow the brain to transform a computer into a component of the nervous system.

Functionally, individual artificial intelligence can be compared to a group of deep learning programs that our brain will create and develop on its own without the involvement of a third-party programmer. This will be possible thanks to the brain-computer interface.

Can a computer system handle the flood of data from billions of living neurons?

The human brain contains about 86 billion neurons, but we are only interested in a small part, about 16 billion in the cerebral cortex. Although technically the interface of the individual artificial intelligence will be able to follow the reaction of a single neuron, in practice such discretion is not necessary. Working with groups of neurons numbering several thousand neurons can provide quite acceptable clarity in signal formation and interpretation.

The brain-computer interface scanner is a gateway and regulator of the amount of incoming data

The most important element of the interface is a scanner that reads information from molecular beacons (binary marker objects) that are injected into brain tissue. The main part of the scanner is a dense array of sensor elements based on specially modified and well-insulated multilayer carbon nanotubes. In the test version, this part can be located on the surface of the epidermis. In the working version, the network of carbon nanotubes will work in the inner layer of the skin (in the fibroblast layer).

The very procedure of installing the scanner will resemble the procedure of applying a tattoo, although it should be much easier in terms of the degree of invasion (injury) since the immune system cells (macrophages) will not be able to “eat” particles of the scanner material.

It is important to understand that the power of the data stream from brain neurons depends not only on the distribution and activity of beacon molecules, but also on the number of signal attachment points and the sampling rate on the scanner. . For correct operation, the scanner must be updated at a frequency of at least 250 Hz, but at critical moments (for example, in a real battle) the frequency may increase due to band narrowing scanner bandwidth. Moreover, the architecture of the system suggests the possibility of modifying these parameters not only to reduce or increase the flow of incoming data, but also to combat distorting artifacts.

Thus, by changing the scanner’s sensitivity or refresh rate, the amount of data will always exactly match the system’s capabilities.


Comments are closed.