In 2029, will humanity be able to overcome the mechanics of machine reason and create an artificial intelligence with human-like intellectual capabilities?

This article explores the possibility of humanity creating a human-like artificial intelligence in 2029. We’ll discuss whether intelligent machines like those depicted in the movie The Terminator are possible, drawing on past and present advances in artificial intelligence to explore the possibilities, including biological approaches.

 

In 2029, The Machines rise from the ashes of a nuclear war (The Nuclear). The Machines wage a fierce war of extermination to wipe out humanity and send their assassins to L.A. in 1984. This is the opening of James Cameron’s movie, The Terminator (1984). The movie is set in the year 2029, but is it really possible for such robots to exist? Is it possible to create machines that look so much like humans and are so physically and intellectually superior that they can dominate humans?
I will try to answer this question in my own way. I think it’s fascinating and fun to think about the possibilities of artificial intelligence and come up with creative solutions to the questions that humans have long wondered about. In the process of conceiving and writing this article, I referenced books by John Searle (Mind) and Haugeland (Aritificial Intelligence). These two books discuss artificial intelligence from a philosophical perspective. I’m going to use them to think about and discuss the problem from an engineering perspective.
Before we think about the Terminator of the future, let’s think about the robots we have today. Today’s mechanical robots have already caught up with humans in many areas. We have robots with better computational and memory capabilities than humans, robots that appear to speak human language, and robots that can mimic many human facial expressions. There’s even research into robots that appear to express human emotions. Now compare these advanced robots of 2012 to the Terminator from the movie. The Terminator is almost human-like in its ability to recognize its surroundings and actively carry out missions. While there are some things that distinguish them from humans, the overall appearance is that of an intelligent creature. In this sense, today’s robots are far removed from the Terminator. Even the most advanced robots of our time pale in comparison to the all-powerful Terminator.
In the past, scientists trying to develop artificial intelligence faced a paradox known as the paradox of mechanical reason. At the heart of the paradox of mechanical reason is the question, “Is it structurally possible for a machine to be smart?” This question arose because it was difficult for people at the time to imagine a machine as anything more than a “tool”. In the past, even the most sophisticated machines were the “sum of simple motions”. For example, the steam locomotive and looms of the Industrial Revolution were machines that changed the quality of life for humans, but their movements were very simple. Therefore, it was difficult to think of them as intelligent. It was as if the words “mechanical” and “smart” were contradictory to each other.
Centuries later, the advent of computers changed this notion. Computers performed computational and memory functions on behalf of humans, and these abilities were clearly seen as attributes of ‘intelligence’. Although controversial, the invention of the computer led some to claim that the mechanics of machine reason had been solved. While there have been many philosophical debates about the question “Are computers smart machines?”, from an engineering perspective, the consensus is that they are. The latest theory of artificial intelligence, connectionist theory, has even shown how to create computers that think in a human-like way. Computers certainly seemed to be getting smarter.
But has the machine rationality paradox been solved? Personally, I don’t think the machine rationality paradox has been solved yet. This is because we have yet to see a human-level AI like the Terminator. Connectionists have given us a direction to solve this problem. But it’s one thing to point the way to solving a problem, but it’s another to actually solve it. In order for a problem to be solved, a complete solution must emerge beyond the idea of how to solve it. In this case, “artificial intelligence” must emerge, and in that sense, we can only say that “machines can be smart” if we can bring a truly smart machine into the world.
The definition of “artificial intelligence” can be ambiguous. I would define “artificial intelligence” as “a machine that can recognize its own environment and react to it in a human-like manner”. This is not a philosophical question, such as “should robots have minds,” but rather an empirical question of whether they actually appear to behave appropriately in their environment.
So, why haven’t humans yet realized AI? To understand this, let’s look at the aforementioned connectionist theory. The founders of connectionist theory, which is currently the most advanced theory of artificial intelligence, were inspired by the idea that the brain is a neural network of countless neurons connected through synapses, and tried to implement artificial intelligence by replicating the brain’s structure through a computer program. The result is the ‘Artificial Neural Network’. An artificial neural network learns by making small adjustments to each neural network node until its output is similar to the output known in the literature. (Actually, the term “learns” can be debated, but at least on the surface, an artificial intelligence network seems to be learning.) If it continues to learn in this way, it can solve even very complex problems. However, no matter how many iterations of training they go through, today’s neural networks are nowhere near the level of artificial intelligence of the Terminator.
There are several problems, but the most fundamental is that our computers are simply not powerful enough to replicate the human brain. Recently, an artificial neural network that recognizes cat faces made headlines. It was said to require 16,000 CPUs as central processing units, even though it had no other function other than recognizing cats. Implementing the entire human brain would require a similar number of CPUs and would eventually hit physical limits.
So, is creating an AI like the Terminator an impossible task for humans? I don’t think so. The physical limitations are the most painful. The way science and technology have been developing, we will eventually run into limits at some point. In order to achieve innovations such as the implementation of A.I., the approach must be accompanied by innovation.
Let’s take a moment to think about past moments of invention and innovation. From the invention of the steam engine, to the development of electricity and the light bulb, to the emergence of the internet, humanity has always been able to innovate by changing the way we think beyond the existing limits. AI is no different, and it requires a change in approach. Rather than simply trying to overcome limitations in the same way, it is important to explore new avenues. One of the new approaches that scientists are exploring is the biological approach. This is the convergence of traditional robotics, computer science, and biology to find new solutions. The idea is that if you can’t solve a problem with purely physical methods, you can take a different route.
The reason we started thinking about biology is because of the functional superiority of living things. There are many examples of life overcoming the limitations of man-made, mechanical, and artificial things. The synthesis of an amino acid chain, which would take two years to synthesize manually, takes only 0.00001 seconds in an E. coli cell. Let’s look at how we can harness the power of these great organisms to implement AI. We can think of three specific examples.
The first is to artificially recreate the DNA sequence that expresses the brain and use it to program AI. The second is to express the brain of a living organism and use it as the brain of a robot. Finally, artificially replicating DNA to create artificial life that resembles existing life seems to be a possibility. One might ask, “Can robots created in this way be called robots in the true sense of the word?” However, if we consider the dictionary meaning of artificial intelligence, there is no problem.
There are also ethical issues that may arise. This is an inevitable consequence of utilizing living things as tools. The idea of utilizing the brain, which is most closely associated with the mind and intelligence, is likely to be met with great resistance. However, if we can overcome these problems and utilize biology well, this approach may be the key to the birth of AI.
So far, we’ve been thinking about the birth of artificial intelligence and the possibility of solving the backward logic of machine reason. Today’s mechanical robots have come a long way from the past, but they cannot actively think and act like humans. This can be seen as a problem inherited from the past, and it seems difficult to solve it in a strict sense. This is because there are great difficulties in physically realizing AI. In this essay, I have proposed a solution to this problem by integrating biology.
Now, let’s return to our original question. In 2029, will humanity be able to solve the mechanics of machine reason and create machines with the same level of intellectual capacity as humans? It’s not a long time to wait for a breakthrough like the rise of artificial intelligence, so it doesn’t seem easy. However, advances in technology and science tend to move rapidly once they pass a certain bottleneck. So we can’t completely rule out the possibility.
If in 2029, we do see the highly advanced artificial intelligence we’ve been waiting for, it may have biology at its core, and this new approach could rewrite the history of science. As technology advances, humanity will face ever more complex and sophisticated questions. In the end, we will have to solve these challenges and open the door to another round of innovation.