Essay: Will AI robots end human civilization?

In early March, a Goh game held in Korea attracted the whole world’s attention. Different from any of the previous Goh games, one of the player, Alpha Go, is an artificial intelligence (AI) robot manufactured by Google instead of a human. Before the game, most of the people, including some senior specialists, thought that Master Sedol Lee who has the Goh experience for over thirty years would defeat the AI robot beyond any doubt. However, the result was really astonishing to the public —— Alpha Go easily overcame Lee by 4-1. A coincidence? Maybe not. The cases that humans are defeated by AI machines can be found everywhere. With the development of technology, especially in the field of computer science and automation, the highly intelligent AI robots which could only be seen in the Hollywood movie in the old days has come into our daily life. It is hard to deny that the appearance of AI robots has make great convenience for the production work and our lives. However, some questions have arisen over whether artificial intelligence will pose a threat to human beings. People begin worrying that the scene in the theater where the AI robots massacre human beings and destroy human civilization will come true in the near future. All of a sudden, the usage of autonomous-weapon system, intelligent robots, supercomputers and many other AI productions are all involved in a heated debate. Famous specialists and scholars, including Stephen Hawking and Bill Gates, have also entered this debate recently.
Artificial Intelligence, which is short for AI, is the name of the field which is defined as “the study and design of intelligent agents”. AI robots are the highly intelligent robots which is made with the technology of AI. There are many famous AI robots that are widely known by people. The movie Terminator and Matrix have a great number of fans all over the world. Other AI movies like Ex Machina and Chappie also perform well at the box office. The science fiction novels sell well in each book store. These science fiction works expand our imaginary space and show us the horrible disaster that might be caused by AI. Other than the virtual AI robots in the movies, there are some famous real AI robots in the history. For example, the AI robot Deep Blue which is produced by IBM defeated the Russian chess master Gary Kasparov in 1997. Another AI robot Watson beat two highest record holder and won over one million bonuses in the TV show Jeopardy, 2011.Too many cases that human specialists lose the games against highly intelligent robots all contribute to the recently panic.
But does it mean that AI robots will really pose a threaten on human beings or even terminate human civilization? Or should we fear the development of AI robots and stop the AI projects? People may be overly worried about it.
Most importantly, AI robots hardly have self-awareness. Self-awareness is the capacity for self-examination or introspection and the ability to recognize oneself as an individual, who is different from the external environment or other individuals. Different from human beings, the intelligence and ability of AI robots depend on the computer program. In other words, what a AI robot can or good at is determined by human instead of itself. The functions and operation systems of AI robots are all designed by humans but not generated by AI robots themselves. The nerve nets and the structures of AI robots are all determined by the input and training data. These nerve nets and structures will help them do well in big data computing and standard operations, but cannot help them be humans. The self-awareness help humans make sure their roles in the whole society so that they can do better in team-work and have greater purpose. The self-awareness also drives humans to achieve more and even rule the world. Contrary to human beings, AI robots may lack the desire to grab the dominions.
The technology has developed a lot from Deep Blue to Alpha Go. Deep Blue defeated the reigning world champion by 4-2 under standard chess tournament time controls. It did a great job calculating every possibility of every single step. However, compared to Deep Blue, which is much like a powerful computing machine, Alpha Go can deal with more complicated situations and have greater learning ability. It seems that Alpha Go can adjust to the changing circumstances and can learn from mistakes, which makes it like a human with self-awareness. However, such abilities are not real intelligence and self-awareness. Instead, it is just the application of simulating human intelligence. All the moves and decisions are designed and made by existing programs and data processing. Though seems bright, the AI robot does not have a physic ontology, which means the whole external world is objective and strange. Different from the complex brain of human beings, the core of AI robots is simplex and specialized, which can hardly interact with the external world. Humans have not rigorously studied the secrets of our own brain, let along apply the relevant discoveries to the production of AI machines. Though scientists are trying hard in programming AI robots into the self-conscious machines in specific situations, there is still a long way for the AI robots to go to realize and understand their roles in the society. Without such understandings, AI robots will never be more useful citizens.
AI robots are, at least for now, assistant tools instead of independent individuals. The applications of artificial intelligence are concentrating on the fields of image recognition, voice recognition, big data computing and industrial production. The jobs AI robots can do is determined right by human beings. Even if AI robots did harm to humans, the perpetrators must be ourselves instead of the AI robots. Instead of end human civilization, they will most likely take the place of many people’s jobs, especially in the field of traditional industries like assembly and customer service.
Even if the AI robots could acquire self-awareness or there were possibilities that they would aim to fight against human beings, we could prevent it by developing an advanced control system or by law and economic approach. The kernels of artificial intelligence are complex algorithms programmed by humans. All the AI robots and AI weapon systems are controlled by the algorithms. Each algorithm is born with its strength and weakness. Advanced control system and related rules are produced to limit the ability of AI robots within a controllable range. Stopping a mad AI robot will be as easy as pressing the off switch of an abnormal domestic appliance. For example, if a AI weapon depending on the GPS system wanted to destroy some big cities in the world, the scientists could stop it by cutting off its network connection to make it misdirected. The same thing is with Alpha Go. Actually, Alpha Go just has a more advanced Goh algorithm than the previous Goh robots. It completes the Goh games within three main steps. The first is to train the policy network and rollout policy by reading data of over 30 thousand Goh games. The second is to enhance the policy network by reinforcement learning and to implement an error-correcting strategy. The third is to generate the position of the next step by using the ordinary network strategy. The core algorithm of Alpha Go is the policy network. To defeat Alpha Go, the opponents can attack its policy network by increasing the search space and destroy its value network.
The legal and economic means are also effective. To avoid the bad influence, Google found an ethical committee in 2014 to guarantee that the technology of artificial intelligence would not be abused. Though the real super AI robots does not exist yet, some rules have already given the direction of the future development of the game rules that AI robots should never harm the overall interests of mankind. Like the laws about cloning technology, some relevant legal institutions can also pass laws in various industries on the use of AI technology. The first is to regulate the production, sale and usage of the AI robots. The second is to establish related legal norms and organizations to do the prior review of AI productions, especially their inner algorithms, to prevent the malicious programs. These laws can prevent the criminals from using the AI technology to make money or commit a crime. Besides, people can avoid the possible risks through economic guidance. For example, people can forbid the production of AI weapons while encourage the production of commercial products. By this way, the production of harmful AI productions can be limited because of the loss of the actual economic value.
Whether today’s AI robots or the super AI robots in the future are dangerous is totally determined by ourselves. Nobody is born to be a thief, so are the AI robots. Everyone hopes that these high-tech products can always service us without harm or any problem. What we should do is to do more research on AI control to specify the production flow and usage criteria of artificial intelligence industry instead of reduce or even stop the research in the field of AI just because the fear of the guesswork.
Still, some people hold the view that AI robots will surpass human beings in each particular field gradually. They claim that though AI robots are applied only in the fields of computing and manufacturing nowadays, these robots with higher efficiency and better concentration have already done much better than human employees. They believe that AI robots have the ability to defeat human competitors in a very short period of time as soon as they step into the same field. They even list some evidence. For example, the data quantity the AI robots can store and process is over billion times more than humans’ brains. AI robots will never get tired or sick as long as there is enough power. The Goh robot Alpha Go can reach the level of a Goh master with the experience over thirty years within just a few months.
However, they may ignore the biggest advantage owned by human: the competence of study and creativity. The synapses in our brains are far more complex and advanced than the logic gates in the core of the AI robots. The AI robots use precise memory address to access their memory for particular information or policies. In contrast, the brains of human beings use the memory which is called the “content-addressable memory”. With the way, the information in our brains can be accessed through “spreading activation” from not only the particular individual concept, but also the closely related concepts. That is why, different from the AI robots which complete the study progress by reading data and making analysis depending on the existing programs, humans are skilled in drawing inferences about other cases from one instance and acting according to the circumstances. After tens of thousands of years’ evolution, human has gained amazing gifts in our genes that cannot be copied by any advanced electronic systems. The sparks in our minds, those unusual but wonderful thinking, shape today’s world and create our civilization. That is why though the AI robots is significantly ahead of humans in the field of automated mechanical production and computing, there is hardly a single invention invented nor a single theory discovered by AI robots.
The true emotions of human beings cannot be simulated by technologies. The abilities to continuous study, create new things and control our emotions make us who we are. Theses abilities can never be truly acquired by Deep Blue, Alpha Go, Microsoft’s Cortana or any other AI robots in the future. For example, when faced with a disaster, even though the AI robots can make the sound of humans’ crying depending on the simulation of humans’ brain, they cannot burst into tears with real sadness. There will never be a robot who can experience the loneliness, happiness, sadness or any other subtle emotions, let alone identify people’s complex feelings. With these emotions, people’s potential and explosiveness are incalculable. You cannot imagine what a mother could do when her children were in danger, nor what a hero could sacrifice to protect the country and the people. Even the day when the war between AI robots and human beings started, we would be the party who have the advantages. As is shown in most AI movies, though humans may be behind the competition at the very beginning, we will laugh last.
In conclusion, AI robots are not heinous gunmen or the harbingers of doom. The scriptwriters and the novelists have overly exaggerated the negative effects of the AI robots. The truth is not as horrible as we might think. Actually, in the game between Alpha Go and Sedol Lee, human beings were defeated by ourselves instead of the AI robots. The human Goh master was defeated by the wisdom of hundreds of scientists and engineers. The final success of highly intelligence AI Goh robots is the success of human’s surpassing ourselves. Instead of feeling afraid of the development of AI robots, we should feel gratified about what we have achieved and how much convenience the application of the technology of artificial intelligence have brought to us. It is understandable that many Goh players and practitioners feel disappointed about the loss against AI robots. However, we should notice that there are more Goh masters who look forward to and are pleased with the progress of AI Goh robots. Some of them even participate in the study of the computer Goh program themselves. In other fields, the future of AI robots is also bright but not horrible. The intelligent traffic system will provide great convenience to people and will reduce the number of traffic accidents. The medical robots will never get tired or sick so that the short of qualified doctors can be solved. The production efficiency of manufacture factories will increase because of the application of robots and intelligent control system. We should be thankful to the talent scientists and engineers without whom the convenience may never step out of the science fiction works and into the real life. There has never been a great invention which never caused debate or panic in out history. The births of telephone, computer and the Internet were all boycotted by a great number of people, including some authorities and specialists. However, without these inventions, we will never achieve the civilization today. The birth of artificial intelligence is not the first event ever which makes people feel afraid and it is surely not the last. Thus, we must put aside the fear and pursue excellence as we always did in out history. We should never rest on our laurels and refuse the application of new technologies. Instead of ending human civilization, the AI robots which represent human being’s wisdom and effort will take our civilization to a new level.