Can AI think like humans?

"Athanassios S. Fokas argues that AI, despite its advancements, is still far from matching human thought, as it lacks the ability to fully replicate the complexity of human cognition, including emotions, creativity, and unconscious processes." (ScitechDaily, Can Artificial Intelligence Think Like a Human?)

Can AI think like humans?

The AI is the biggest revolution in history. That means it comes to stay. And even if we hate that thing, we must start to live with it. When AI generates answers, it will use the Internet as a database. How well the AI can connect and search data and form the network determines its accuracy.

In the past, AI was a chess program. And researchers used those programs to test their algorithms. People thought that humans were, invincible until Deep Blue AI won over chess master Garry Gasparov. The fact is that Gasparov won the series. But the Deep Blue won a couple of matches. There have been lots of advances in AI since those days. That happened in 1997.

Today there are lots of more powerful computers and the advanced Internet that allows computers to drive heavier and more complicated code than ever before.

Today AI is much more than some chess computers. Chess is an easy thing to model for computers. The game area is limited, and buttons have straight regulations. But if we want to make robots that operate in everyday missions, robots must have more complicated programs than some chess programs.

When AI selects data sources, it uses certain algorithms for making that selection. The AI doesn't know what reads on those home pages. But it also can compare information about the internal sections of the home pages with other sources with similar topics. So when the AI makes mistakes, those mistakes are in sources.

But are we afraid the AI because it's too perfect?

Do we fight against AI, because it's dangerous? Or are we just afraid of it, because we think of that thing as a competitor?

The AI is not perfect. It's more effective as a coder than humans. The AI-controlled drone can resist stronger G-forces than manned aircraft. And the AI has no feelings. That means it is never angry. It's never sad, and things like public opinion do not affect AI and its decisions. Or if programmers want the AI can let people vote about the solutions that it has.

"New research addresses the risks and liabilities associated with implementing AI in the food industry, proposing a temporary adoption phase to assess AI’s benefits and challenges, and emphasizes the need for more research on legal and economic structures." (ScitechDaily, The Paradox of Perfection: Can AI Be Too Good To Use?)

But does AI think?

The fact is that the AI is not yet thinking. And if we want to make the AI, that thinks like humans, we have two choices.

1) We can make database solutions where there are billions of databases. The human brain involves about 100 billion neurons. The synapse connections increase the number of those neurons.

Interconnecting neurons that system can create virtual neurons and make database combinations. That ability makes it possible. That there could be more than 250 billion combinations of the databases.

So AI that thinks like humans requires 100 billion microchips with at least 100 billion databases, that this system can connect.

2) We can make a computer that uses living neurons. That kind of hybrid computer is more effective than before.

Image: Deep Blue versus Garry Kasparov - Wikipedia

In normal cases, the AI is a language application that turns human commands into models. That computer understands. Researchers can put this kind of language model into the microchip kernel. And that thing makes it possible for the next-generation computers can follow spoken commands.

As I wrote many times before in those applications, the language model plays a key role. It is at the center of all applications. That language model is used to command the computer in the center of other applications.

The language model can use a speech-to-text application to drive spoken commands to the application. That kind of AI has a model where it can search for safe and confirmed information. There is a list of the file descriptions that deliver trusted and confirmed data. Those things are governmental and university homepages.

After that, the AI collects homepages and connects them to texts and data in a certain order. There is also a model of how the AI can make correct text that has a good vocabulary. So the system knows predicates, subjects, and other things. That is required to make fluent text.

But some other types of AI react to situations. That kind of AI tool sees that something happens it comples that thing with things. That is stored in its memory. If there is a match the AI makes the act. That connected to that action. These kinds of AIs are things that used to take thieves. Or the AI, that controls drones can make escape and evasion movements when it sees incoming missiles or AA-flak.

The AI seems like thinking. But it just collects information from the Internet. And then it connects that information. The AI can make the answer with very high accuracy. And how good-looking answers the AI can give depends on the data mass that it can use. Also, things like polite answers are making AI seem human.

If we think that the AI-controlled robot is programmed to answer to puch with puch, that thing makes it human. The robot can also say that it hurts when its sensors detect a punch. That is hard enough. The robot can also call the police if some vehicle hits it. But those robots don't think.

Thinking AI requires the living neurons. Neuro-computers that use living neurons can think. The biocomputer always thinks. But those neurons require information that they connect. Thinking is simple connecting databases and making the new networked entireties about them.

When we think that some AI turns against humans, we might use things like EMP weapons against it. The EMP weapons are usually quite harmless to people, but they destroy the electric components. AI is a tool that will make many things different than before. But it can serve humans.

The thing with AI is that we must not let AI think for us. When we talk about military applications. We must realize that modern warfare is more complicated and changing than ever before.

Things like GPS-guided bombs are predicted, to be easy to jam. But reality is another, and at least Russians have not successfully cut the GPS signals. In the Western world, we think, that the military should protect the land and its people. We must realize that the military is a brutal world. Its purpose is afraid of enemies. And then we must realize, that what we define is the solution that we make good or bad.

https://scitechdaily.com/can-artificial-intelligence-think-like-a-human/

https://scitechdaily.com/the-paradox-of-perfection-can-ai-be-too-good-to-use/

https://learningmachines9.wordpress.com/2024/01/23/can-ai-think-like-humans/