How Human Beings Face The Challenge Of AI

Zheng Lei / Wen

A major concern of economists is: what kind of impact will artificial intelligence bring to employment? A consensus seems to have been that artificial intelligence will temporarily impact the job market, resulting in a sharp increase in structural unemployment. But in the long run, because of the opening up of new technologies in the new economic field, this unemployment problem will not necessarily be more serious than it is now. Thomas Davenport, the father of data analysis in the field of artificial intelligence, does not think so. In fact, the impact of AI on employment is very similar to industrial transfer, and we have all seen the social consequences of the hollowing out of manufacturing in the US and Japan. In his masterpiece "human-machine symbiosis: the 5 great strategies of human beings in the age of intelligence", the focus is on how knowledge workers can achieve greater competitive advantages than intelligent machines in the AI era. His suggestions are insightful. Of course, this is not necessarily the only way for human beings to go.

As long as AI is invading a certain field, human job opportunities in this field will be reduced. Looking back at the facts of AI development, we can see that the first stage is to liberate people from those exhausting jobs. For example, Foxconn's electronic product assembly work, Amazon warehouse selection and packaging work, gradually replaced by the robot. In the second stage of automation, boring, tedious and transactional work will become the target to be replaced. Most of the daily jobs of white-collar workers belong to this category, such as secretaries, assistants and so on. These stylized things can easily be converted into computer executable code. It can be said that in the first two stages of automation, the economic structure based on simple manufacturing and general services will be eroded, causing a large number of laborers who need low skilled jobs to lose their jobs. As machines become more and more intelligent, they can also accomplish some decisions that we once thought only humans could do, but machines could not be competent. At present, applications of artificial intelligence, such as chess playing and examination machines, have proved that intelligent machines can make decisions faster and better than human beings.

A survey of 352 machine learning researchers shows that AI will surpass human intelligence in 30 years. In the next 5-10 years, intelligent machines will work with people in most industries. McKinsey predicts that the drastic change of knowledge workers' work will be replaced by automation in 2025, when intelligent machines can complete the work of 110 million -1.4 full-time employees. Davenport believes that in developed economies, about 25%-50%, who are mainly engaged in knowledge and information processing, will face the threat of AI. In the United States, this is equivalent to about 47% of the jobs.

When AI is intruding into the field of knowledge work, some jobs no longer need the necessary skills before, and half technology or unskilled workers can accomplish these tasks. This phenomenon is called "deskilling", and this prospect is quite unexpected. For example, the rich imaging diagnostic ability accumulated by the radiologists for many years will no longer be needed. Now the technology of machine reading CT and MRI images has emerged. The machine has already marked the lesion site, and the number of these doctors has been decreasing continuously in recent years.

If artificial intelligence is regarded as a "positive elimination" instead of low end labor, partial or even complete substitution of knowledge workers should be regarded as "Reverse Elimination". Although both of them are revolutionary in purely economic and technological terms, they both have a great impact on social structure and normal social life. For technology, we should not only consider the advanced nature of technology, but also consider the justification of its means and consequences. The application of AI technology should aim at enhancing social welfare and realizing people's inner happiness. From a certain point of view, human beings have many shortcomings, such as higher labor costs, providing ergonomic working environment and facilities, and logistics support, which are more expensive. Human workers also need extra supervision, such as fraud or dereliction of duty (according to the US human resources manager survey, only about 13% of employees are working hard at their jobs). This is much more expensive than using the same intelligent machine. However, labor has brought an indispensable sense of belonging, satisfaction and sense of achievement to human beings, which is the most valued identity and core values of human beings. Even if no working people can share the wealth created by AI system, such a society can not let everyone reflect their value and get the meaning of life. For mankind, this is a very cruel prospect.

Davenport pointed out that the conditions for the harmonious co-existence of human beings and intelligent machines are to make themselves the global, the avoidance, the participants, the experts and the pioneers. I am afraid that individuals who can achieve this standard are scarce. Human beings must dominate the development of artificial intelligence, which is the inherent requirement of human society. The orientation of AI can only supplement human's imperfect action ability. Any technology that infringes human's labor rights from efficiency alone is against ethics and is not worth advocating and encouraging. From this principle, we should be cautious about the application of AI technology with the aim of simply improving efficiency.