The rapid development of technology plays a cruel joke on us. Technologies have developed so much that their accumulation creates a feeling of helplessness in people. It is impossible to fully understand how this or that device works. An endless number of features, bugs, hidden settings and menu items makes you think about the finite resources of the human brain and human life.
On the one hand, smart machines and computers are created to make human life easier and better. On the other hand, a person often spends too much time and attention on all this diversity.
Let’s define the concepts
Digitalization, artificial intelligence, bots, neural networks – we hear these words too often. So often that it’s time to figure it out and comprehensively evaluate whether it’s good or bad?
As usual, let’s first define the terms. What characteristics can artificial intelligence have?
- Collection, accumulation and processing of information.
- Self-study.
- Distribution of information to the outside world.
- Communication with a person.
- Making decisions. Perhaps this is the most important point.
These items may include systems such as a machines, robots, computers, production equipment, and so on. They all have some semblance of intelligence in one way or another. And have one or more of the characteristics described above.
But any artificial intelligence has one distinctive feature – it is created by man. This means that it was man who endowed him with certain abilities and parameters.
Well, have you figured out the terminology? Me not. The fact is that now the boundaries of various concepts are very blurred, and it is difficult to give a clear definition of one or another system. A typical example that is always at hand is a telephone. What is it – a computer, a self-learning device, a means of communication and entertainment, a means of controlling a person, artificial intelligence, a device for digitalization?
A few words about digitalization
The advantages of digitalization are obvious: it facilitates human labour in organizing and processing information. In our country, digitalization is often understood as the maximum transfer of human life to “digital tracks.” This includes payment cards, electronic documents, digital passports, biometrics, control of movement and access to facilities.
There are already volunteers who have had a chip implanted under their skin, which carries all the information about the owner – from his passport data and location to information about his health status and details for payment at the supermarket checkout. Comfortable? Maybe.
Digitalization works great as long as everything goes as it should. If you are a law-abiding citizen, a cog in the system, then you won’t even think about breaking or changing anything.
Two quotes on topic:
“The best slave is the one who thinks he is free.”
“I’ve seen too much of the world. I’m a bad slave.”
But here it can be brought to the point that the line between a person and a robot will become almost intangible – both will obediently carry out the will of the “master”. There are more than enough films on this topic.
Perhaps the best example of artificial intelligence is neural networks. The main property of a neural network is the ability to quickly learn. A person only sets the desired vector at the initial stage, and the neural network continues on its own, exponentially increasing its knowledge and, dare I say it, creative abilities.
For example, a neural network creates texts that are almost indistinguishable from texts written by a person. An example is the YaLM (Yet another Language Model) language model from Yandex, on which Alice works.
There are other services, for example, for autocomplete sites. This unpleasant phenomenon is gaining momentum in RuNet. The neural network takes several articles on a given topic, and in a few seconds generates quite adequate text with high uniqueness. As a blogger who writes articles with his hands and head, this phenomenon is extremely unpleasant for me, but I don’t know how to deal with it.
Another example is neural networks that create images. Just set a text description and choose a style, and you can get a very interesting image. Example of an image generated for the request “electrical market”:
How neural networks are introduced into humans
I have already talked about chips implanted under human skin (an invasive method). Here it is more correct to talk not about neural networks, but about neural interfaces. But what prevents the neural interfaces of many people from being combined into a single neural network?
Initially, such technologies were developed to significantly improve the lives of people with disabilities. For example, paralyzed people can control cybernetic prostheses, computers and other gadgets in this way.
Further more. Nowadays, many companies have mastered technology that allows them to write texts and communicate on social networks without using a keyboard, using, so to speak, the power of thought. Of course, this method of communication makes it possible to play computer games.
The most famous company in this area is Elon Musk’s Neuralink. Her successes in the field of neural interface are amazing. Neuralink managed to move further than its competitors did.
Now a completely wireless chip implanted under the skin is being tested on monkeys. But I’m sure people are also participating in the trials.
The main thing is that Neuralink technology allows you to transmit information directly to the human brain. For people, installing such a chip makes it possible to restore vision. Even for those who have not had the opportunity to see since birth. After all, the visual part of the cerebral cortex is still functioning. The same applies to people with hearing loss.
In the future, a completely immobilized person will be able to find a full life. After all, information can be stimulated and transmitted to any part of the brain. And the artificial reality received in the brain can be augmented, embellished, changed. For a neural network there is freedom here!
An example of such a future is shown in the film “Surrogates”, when people lay at home, and instead of them humanoid android robots lived a full life.
Spoiler: the main villain who created these androids united them into a neural network and thus gained power over all of humanity. But then Bruce Willis, as usual, saved everyone and returned everything “back”.
Soon it will be difficult to distinguish a person from a robot
I hope I conveyed to the reader the main idea of the article – any artificial intelligence endowed with logical thinking and the ability to create and make decisions can be harmful and dangerous to humans. It is even more dangerous when such intelligence takes the form of a neural network capable of learning.
Criticizing? I offer! To ensure that the neural network does not pose a danger, it must be limited by the following rules:
- The operation of any system with artificial intelligence can be terminated at any time without harm to the person and the system. That is, in one form or another there should be an easily accessible “Emergency Stop” or “Cancel” button. For example, it is necessary that the robot can be turned off in any situation, and the use of the neural interface can be easily stopped.
- A person should be able to immediately understand that they are dealing with a robot. That is, any artificial intelligence should immediately make it clear to a person what he is dealing with. As with the law on advertising – looking at any text or image, we must immediately understand that this is advertising.
This is perhaps the most important law, especially relevant today. I would even put it as the first law of robotics:
“The first thing a robot should do when meeting a human is to identify itself as a robot.”
The laws of robotics need to change
The science fiction prophet of Russian-Belarusian origin Isaac Asimov proposed the famous three laws of robotics, which would be logical to mention in the context of this article. These are the laws in my interpretation:
1. The robot must take care of human safety.
2. A robot must obey human orders unless this contradicts the first law.
3. A robot must take care of its own safety, unless this contradicts the first and second laws.In my opinion, these laws are extremely vague. Both the robot and the person can interpret them in any way they want, depending on their understanding of the situation and immediate benefit. Isaac Asimov himself repeatedly raises the topic of these contradictions in his stories.
Many films have been created on this topic, in which artificial intelligence makes decisions that are disastrous for humans or even for humanity as a whole. And the robot makes such a decision for a simple and logical reason – it begins to consider a person a dangerous creature.
The main thing is that humanity must apply the laws of robotics first of all to itself. Of course, this seems like a utopia, but if it happened, the world would definitely be a better place.
