5 myths about artificial intelligence that are far from reality

Myth No. 1. Artificial intelligence surpasses human intelligence

The name itself, “artificial intelligence,” may suggest this idea. It sounds as if we are talking about a complex technical mechanism, reminiscent of the brain. The growing popularity of visual neural networks in recent years adds to the illusion. The generated pictures are scattered into memes and hit headlines in the spirit of “The neural network showed what Winnie the Pooh will look like in steampunk style.” Because of this, one may get the impression that artificial intelligence itself is engaged in creativity, sets tasks for itself and shares the result.

In reality, AI is simply an algorithm designed to solve specific problems: answer calls, select music, or draw pictures. In order for a neural network to “show” something, a person is needed to enter a request into the system. In addition, artificial intelligence is only capable of copying existing solutions, but not creating them. This is why AI is good at tasks like “draw superheroes in the style of Van Gogh” - both key parameters of the request are known and have many references on the Internet. And a request in the spirit of “draw something original and unlike anything else” will most likely confuse the neural network.

Myth No. 2. Neural networks know everything

Artificial intelligence knows as much information as the developers “fed” it. He can also use data that is publicly available. These are not only current materials, but also outdated web versions of newspapers and magazines, numerous forums and comments.

You can verify this yourself. If you enter the same query into conditional ChatGPT, then most likely the AI ​​will give the same answer every time. This is why the profession of industrial engineer is so popular today. Such specialists are able to give the correct hints to neural networks and actually speak the same language with the AI, which helps to obtain the desired result.

Myth No. 3. AI can make decisions like a human

Smart systems are now used in many areas, from law to agriculture. Even music and films are already selected for us by algorithms. But there is no reason to panic. Take medicine, for example. The specialist here does not fall out of the diagnostic process; AI simply allows him to rely not only on professional experience, but also on big data analysis. Neural networks can quickly study thousands of medical records, find patterns in them and suggest what kind of deviation a particular patient has. Or assess the patient’s condition using such non-obvious signs as, for example, sound of voice. In this case, any important decisions regarding treatment are made by the person.

Another example is traffic cameras. They are loaded with a simple algorithm that compares the driver’s behavior with the requirements established for a specific section of the road. But technology can make mistakes. For example, you won’t be able to see a fastened belt if it blends in with your clothes. In this case, the driver retains the right to appeal the fine, and this is already dealt with by specialists. That is, the functions of control and making the final decision are still assigned to the person.

Myth No. 4. Technology will take jobs away from people

By automating one job, technology always offers another. For example, when the first computers began to appear in companies, typewriters were no longer needed. However, the world is not overflowing with thousands of secretaries embittered by complex equipment. Specialists were retrained and mastered new skills. In addition, AI has limitations that prevent its full expansion into the labor market. First of all, it is a focus on completing a specific task. Algorithms are not able to act flexibly and creatively, to assess sudden circumstances and force majeure.

Perhaps in the future some professions will transform. For example, in 2020 in Russia came out the first book officially translated using Yandex Translator. The service processed 350 pages of text in 40 seconds. But this does not mean that learning a foreign language is now a hopeless skill. It’s just that the translator’s function is shifting towards editing machine text. People will remain in demand in the field of fiction. A machine will most likely never be able to process prose or poetry in such a way that the author’s style is preserved.

Myth 5. Smart devices and robots will enslave humanity

Hollywood movie scripts are impressive. And they give rise to misconceptions: they say, over time, technologies will begin to improve themselves and will succeed in this so much that a person will lose control over them. To dispel the myth, it is enough to remember that neural networks are just databases with a bunch of formulas that were created to help people. They are devoid of emotions, so they are not capable of suddenly wanting to destroy civilization.

Some modern algorithms are capable of learning. For example, an autopilot in a car can capture information about its environment and use it to improve its responses in the future. But a truly self-aware neural network is still something out of science fiction. Even if it seems to you that the voice assistant is sincerely laughing at your jokes, you should not delude yourself. His “emotions” are the result of an algorithm that contains certain patterns of human reactions.