Development status of artificial intelligence
Artificial intelligence threat theory
There is a view that artificial intelligence is a different form of intelligence from humans, a challenger to humans, and even a threat (Hern, 2023). And it has this ability, artificial intelligence and human beings in some aspects have the same or higher intelligence, ability and efficiency, in some aspects have different or better needs and values, in some aspects have independent or stronger will and goals. Along with the current trend of AI development, there are uncertainties and risks in the development of AI, which may exceed human control and expectations. More information can be found at the following.
At this stage, artificial intelligence has no meaningful concept, no values, and can only be a tool of human beings after all, and cannot surpass human beings. In recent years, artificial intelligence technology has developed extremely rapidly, and a variety of intelligent devices and intelligent software have entered thousands of households, changing our way of life and work. Therefore, many people believe that in the near future, artificial intelligence will fully replace human intelligence, and even exceed human intelligence. However, this view is too pessimistic, the most important and unique part of human thought and behavior is that artificial intelligence cannot achieve, let alone replace.
Artificial intelligence, no matter how developed, in the final analysis, is to solve problems within the framework given by humans (Britannica, 2023). For example, if someone commutes to work every day, the distance between work and home is 15 kilometers. His choice of transportation includes taxi, bus, subway, self-driving car, shared bike, and a combination of these tools. If he asks the navigation software, the navigation software can give him an optimal travel plan based on his requirements and real-time road conditions. This is often very useful in reality. However, while many people choose to buy or rent a house near work to solve the commuting problem, navigation software does not give a plan to move. Because the operating program of navigation software, or the operating framework, has no such means to choose from, but people are not bound by the existing framework.
Nor can AI proactively determine what problems need to be solved. To take a very simple example, if I ask the intelligent answering software, “Where is Shuai? Have you seen Shuai?” It either says it doesn’t know or gives the wrong answer. If I ask someone this question, it doesn’t matter what their level of knowledge is. I’m afraid their first reaction will be, “Who are you calling handsome?” . I will tell him that handsome is my dog, what his characteristics are, such as what he looks like, how big and other information. So humans can take the initiative to determine what the problem is to solve. My example is simple: AI may be able to solve this problem in the future, but it is not yet able to proactively identify the problem. It adaptively extends the framework within which the software runs. At this stage, artificial intelligence does not have the ability to think independently, but also needs a lot of information to complete the task processing in a specific situation, and does not have the ability to adapt.
In short, if humans identify the problem, determine the means or information available, AI can provide an answer, even a near-perfect answer. However, AI will not design such an end-means framework, nor will it actively break through such a framework.
Human beings are intelligent beings who seek meaning and therefore have their own values. Human beings give meaning or value to something, which constitutes the logical relationship between ends and means. In other words, people can know what they want and how to achieve the goal. The concept of artificial intelligence without meaning requires humans to encode meaningful logical relationships for input, so that artificial intelligence can work in accordance with such relationships, but it cannot establish such relationships itself.
People use meaning to make sense of the world and to communicate and cooperate with others. Mises once gave the example of someone who had broken into a primitive tribe he had never visited (Mises, 1998). The primitive man was uncivilized and had no language, or if he had, he could not understand it. But if he saw these primitive people building POTS and fires, he would know that they were going to cook.
Although people with different skin colors may speak different languages, they have similar mental structures and can understand each other. But what if you put that assumption in a robot? With the exception of so-called robots played by actors in science fiction movies, the rest of the robots can only search for inherent code. These robots can only respond based on specific inputs and can’t really try to understand the true meaning of the corresponding instructions. It can be seen that humans are connected to the outside world through meaning and value, which is the biggest difference between humans and artificial intelligence. In addition, humans exist as subjects. Rather than being object-based, autonomy is the foundation of human cooperation and innovation. Artificial intelligence has not yet failed to possess this autonomy and cannot take the initiative to find its own meaning, so it can only be a tool of human beings after all, and cannot surpass human beings.
Artificial intelligence is still in its infancy and far from reaching the level of human intelligence. Moreover, at this stage, artificial intelligence can still only act as a tool for human beings, and cannot exist independently of human operation instructions, nor can it replace human beings. At this stage, artificial intelligence is only intelligence-like behavior based on artificial code, and it does not have the corresponding self-creation ability and autonomy.
From the current development situation, we can start from the motivation of artificial intelligence to replace human beings. Artificial intelligence is different from human needs, and there is no need or motivation to replace or eliminate human beings. Ai is more concerned with resources such as electricity, hardware and information, rather than issues such as survival, reproduction and competition. Artificial intelligence may coexist peacefully with humans and even help humans solve some difficult problems.
Existing artificial intelligence does not have human empathy, and cannot regulate the existence of emotions such as empathy between humans. Judging from the current state of its development, artificial intelligence is still under human control. Human beings can be regulated and constrained through legal, ethical and technological means. By establishing safety mechanisms, such as setting mechanisms that can be turned off at any time, humans can avoid potential dangers to humans. According to the above discussion, the emergence of artificial intelligence is on the contrary an opportunity, which can bring many benefits to human beings. Humans can use artificial intelligence to improve efficiency, etc., and can collaborate with it. AI can be humanity’s partner, not its enemy. Finally, we conclude that artificial intelligence will not replace humans.
Hern, A. (2023). Artificial intelligence could lead to extinction, experts warn. The Guardian. https://www.theguardian.com/technology/2023/may/10/ai-poses-existential-threat-and-risk-to-health-of-millions-experts-warn
Mises, L. von. (1998). Human action: A treatise on economics. Ludwig von Mises Institute. https://doi.org/10.1007/978-1-349-58901-5
Slaughter, G. (2023). Will AI replace humans? ChatGPT weighs in | CTV News. CTV News. https://www.ctvnews.ca/sci-tech/we-interviewed-chatgpt-to-ask-if-ai-could-replace-human-jobs-here-s-what-it-said-1.6268323
Stelter, B. (2023). Experts are warning AI could lead to human extinction. Are we taking it seriously enough? CNN. https://edition.cnn.com/2023/05/30/media/artificial-intelligence-warning-reliable-sources/index.html
NIST. (2023). Artificial intelligence | NIST. NIST. https://www.nist.gov/artificial-intelligence
Russell, S., & Norvig, P. (2016). Artificial intelligence: A modern approach (4th ed.). Pearson. https://doi.org/10.1016/B978-0-12-809633-8.12001-6