Artificial intelligence (AI) is a new technology to support human work on multiple areas, it could also replace humans in many physical labor in society with a mechanical arm or body. AI advocates artificial intelligence, which is the intelligence exhibited by machines made by humans, which has similar functions with human minds such as learning. There are many discussion points about whether AI replacing human labor is beneficial to society or not. First, in recent years, AI technology is still not stable and economical in most industries, which costs too many resources for few results. Furthermore, as a created program, AI cannot think like a human, which means it does not have human nature and does not judge its capabilities on its own. However, AI is beneficial for repetitive jobs such as item handling and sorting, which has increased the efficiency of some easy, high-repetition missions. Finally, the artificial intelligence program has not been successful in recent times, further technological development would provide more possibilities and potential for the development of artificial intelligence programs. Say no to plagiarism. Get a tailor-made essay on "Why Violent Video Games Shouldn't Be Banned"? Get an Original Essay First of all, artificial intelligence programs are still in development. Basically, in recent years, artificial intelligence is still a semi-finished product, the hard work of humans has not helped the program's artificial intelligence. Most released AI programs still follow orders and calculate all possible steps, then start work, there is no process for AI “thinking” or “learning”, because those, AI is not completed. For example, AI-based driving is a popular word in society. It is normally known that a car receives information about the environment from a variety of sensors it has, such as GPS, radar and sonar.(2) Therefore, the AI ​​program needs to input information into the developed system, and then calculates each possible for the different movements and sends the steps to the mechanical movement. He still followed orders in the process and spent time transporting information between sensors, main computer and mechanical part, which are processes in excess of a human driver. Another example is the Internet assistant, Apple Co.'s Siri is a famous product. Siri is the most common voice assistant program on Apple phones, which uses a microphone to receive the voice and uses the Internet database to find the best answer, then respond to the user with designed phrases. However, the assistant program cannot work without the Internet, which is the database, also it may misunderstand the order of users or recognize multiple languages ​​in one sentence such as the English character E and the pronunciation of One (Yi) in Chinese. There is no "learning" function of the product yet, except recording the user's daily information and designed suggestions. Not only above, the artificial intelligence program normally needs a supercomputer with high-speed computing capabilities to help them collect data and choose the best answer, which has volume as much as a class, also needs huge amounts of resources for work and maintenance. This is a high-risk investment before the AI ​​program does some work worth the machine's worth.it is based. Therefore, the artificial intelligence program is not ready to become a major part of human life, it is still incomplete, and it is not affordable enough to be accepted by people. Furthermore, AI programs have problems with “thinking”. As programs, they have no ability to judge what is right or wrong unless the developer gives them an exact answer. Even though they have a large database as support, some easy-to-read questions still led them to misunderstand. The main problem is that the AI ​​programs we have are not like we thought, that is, they could “think” like a human. Pronoun disambiguation is an example of how AI “thinks” about the problem, such as “City councilors refused protesters a permit because they feared violence. What does the word “they” refer to here? “Humans could make decisions easily and quickly, which AI did is “just a little better than random.” Because natural language is also difficult to understandmeaning in different environments or backgrounds, a program designed like artificial intelligence is difficult to recognize well, which human could simply do. There have also been problems with using AI as a lawyer. Special programs can have greater efficiency in censoring case information and predict the possible judgment, however, I am still in trouble with cases involving the interpretation of the law, such as "did someone perform poorly because they spent too much time watching "Wait?" or “do you give a wrong judgment on purpose or not?” These philosophical components will interfere with AI judgments. In fact, not only will philosophical problems disturb the AI, but also biased information would lead the AI ​​to make a wrong judgment, which means that the AI ​​is not reliable and correct enough without real information input. Therefore, AI programs cannot replace some human works on society because they are not "human" enough, which cannot think or feel like a real person, also they will make mistakes that humans will not make. On the other hand, artificial intelligence programs are also still being developed and completed, they are still getting benefits in some areas with simple order but hard jobs, which are reflected as robots. With a mechanical arm or body, these AI robots could carry heavy objects or multiple pieces of material. They could also protect humans in different working environments. In 2016, Amazon deployed approximately 45,000 robots in its distribution centers and warehouses. These robots could pick up and carry objects weighing more than 340 kg. With the help of robots, Amazon warehouses greatly reduce the time between ordering and shipping to the warehouse. As experts presented, robots replacing human labor is a “general trend”, with automation developed, more and more robots have been used in society, such as Amazon's warehouse robots to transport and transfer items, which increased the efficiency of warehouse picking and classification. Another example is automotive manufacturing robots, which usually have a long metal arm for assembly work such as spot welding and painting. They have improved quality and reduced warranty costs, plus they protect workers from dirty, difficult and dangerous jobs. Robots will not drop parts or handle them in a way that causes damage. They reduce waste previously caused by human errors.., 114, 254-280.