Artificial Intelligence and its Use in IT

This article will not use information from Wikipedia or other sites, nor will it use the generally accepted marketing PR of "artificial intelligence". Let's consider only an objective view of the essence of AI and its practical application, without beautiful words, only a description accessible to everyone.

What is "Artificial Intelligence"?

It is worth starting with the most important thing, with the absence of intelligence as such. This is the biggest misconception of most people. AI is not endowed with reason, despite beautiful marketing. Having analyzed even the description of AI in Wikipedia, you can understand that the term itself arose long before the advent of neural networks and other things, back in the days of calculators. But since the world is ruled by money, marketing got involved and turned the ordinary into something unique.

"Artificial intelligence" is an array of algorithms that determine its operation. And its distinctive feature from calculators and computers is that the former have only one answer to the question posed, one option for solving the problem. While AI has a huge number of "or", that is, answer options. All answers are obtained from the Internet, by reading sites and collecting all the information in one database. The AI considers the most correct answer to be the one that occurs most often. This is not the most truthful information, not the most verified, not the most reliable, just the most popular.

For What Purposes Should "Artificial Intelligence" Be Used?

The so-called AI has practical applications, but only in the field of static data. It cannot be used to analyze dynamic data. What do we mean by "static data"? It is any data that does not require logical analysis. AI can save us time that we would spend reading a large number of sites, since it has already read them all and knows the most likely answer. But even when processing static data, you need to know how to use AI. That's right, I did not misspeak; perhaps the most advertised product in the world still needs to be used. Below I will show you one example that will demonstrate what I mean.

Let's enter a phrase into Google's search bar and see what its AI tells us (AI Overview).

• Is Disodium EDTA safe?
• Short answer: Yes, it is considered completely safe and is not considered toxic.

Now let's just change the search phrase slightly.

• Disodium EDTA dangers
• The answer now sounds completely different: it is dangerous and can damage the kidneys.

The example chosen is not the most frightening, so as not to scare you too much. Approximately the same variation of answers can be obtained for most queries. For example, substitute Ceteareth-20 and you will get a variation from safe to carcinogen.

How AI Works

The analysis of chemical safety clearly demonstrates how easily AI misleads us. This is not done intentionally; such a variation of answers is obtained due to the principles of operation of "artificial intelligence". It gives you in response exactly what you are looking for. The main difference between the mind and AI is that the mind analyzes the received information, compares information from different sources. Our mind tries to make patterns, create connections that make this or that result possible. The human brain creates new words that are not in any of the articles read; makes conclusions that are not found anywhere. Simply put, data analysis occurs in our heads. AI, in turn, searches for what you asked and gives the most probable answer only if the requested one matches. AI is not capable of drawing a conclusion from information that is not there and to which the article does not lead, since it does not have intelligence. This is only a probability theory, which is based on a huge database of information.

How to Work with "Artificial Intelligence" in the It Sphere?

In my opinion, it is impossible to use AI to solve programming problems or use it as a system administrator. Moreover, it is very harmful. The efficiency of such an "assistant" is no more than 10-15%. On the one hand, we can say that this is a 15% improvement in work, but on the other hand, this is 85% of errors that can lead to fatal consequences.

Conclusion: Is AI Worth Using?

AI is definitely worth using to speed up the search for answers to questions that do not contain variables that do not require analysis. But even in this case, the result provided to you by the AI assistant is worth double-checking by approaching the question from a different angle and radically changing the search phrase.