What major does the university study so that it will not be replaced by AI？

　　Volunteering to fill in the report is more brain-burning than the college entrance examination itself. Majors, colleges and cities all affect the candidates’ trend in the next few years or even their lives. In today’s era of rapid depreciation of diplomas, students will not only play games with their peers, but also deal with a new opponent — — Artificial intelligence.

　　For example, in May this year, GPT-4 was reported to have passed the four major accounting industry qualification examinations, with an average score of 85.1. This has brought a lot of anxiety to people who are studying. Some people commented under the relevant reports: "Is it going to be replaced by the painstaking examination?"

　　This is not alarmist, computer, translation, law … … Search the positions corresponding to these majors with AI as keywords, and you will seeMany professionals have expressed their concern about being replaced by AI, and some have even encountered the problem of being transferred or even unemployed because of AI.

　　How to choose a major to avoid the impact of AI? What kind of job is less likely to be replaced by AI in the future?

　　In March this year, Goldman Sachs released a report on the impact of artificial intelligence on employment, "The Potential Large Effects of Artificial Intelligence on Economic Growth", in which 13 out of 39 job activities were judged as AI competent. Based on this analytical framework, we evaluated and calculated the AI substitution index for nearly 20,000 tasks in more than 900 jobs in the O*NET database.

　　It turns out that,The tasks that can be replaced by AI account for over 60% of the jobs, accounting for 31.1% of all non-manual jobs, but the jobs that AI can be 100% competent are not available yet.

　　How will artificial intelligence affect or even reshape our work? Should we actively respond and embrace? Or should we adopt a contempt strategy and avoid self-disorder? The answers to these questions not only depend on the development and application of current AI technology, but also may be hidden in the past. Since the industrial revolution, how has technology shaped careers? This history may give us some inspiration when we encounter technical singularity.

　　Electrical and computer … … AI is not the first new technology that scares human beings.

　　On May 2 this year, tens of thousands of Hollywood writers held a strike protest. Different from the past, this time they pointed their finger at AI— — It puts forward the requirements of refusing to provide changes to the AI draft and refusing to add AI-generated content to film and television works. Screenwriters worry that AI not only takes away their jobs, but also devalues "the dignity and value of human labor".

　　Although this is the first time that human beings have gone on strike for boycotting AI, it is not the first time that human beings have boycotted new technologies.In the industrial revolutions of more than 200 years, textile machines, steam engines and automatic telephones … … These tools, which seem common now, have caused panic and fierce resistance when they first appeared.

　　In 1965, when computers were first used in offices, American writer eric hoffer wrote a warning in The New York Times: "A group of skilled Americans were deprived of meaning and value." This statement is exactly the same as the reason why Hollywood writers protest AI today.

　　And the logic behind this is easy to understand. Electric lights make the city’s lamp drivers no longer needed, and the operators are redundant because of automatic telephones. "If something threatens a person’s livelihood, it is only natural to resist it.For most urban residents, skills are their capital, and they are able to support their families because of these human capital. Karl B Frey, a researcher at Oxford University, wrote in the book "Technical Trap".

　　But if you think that the emergence of new technologies will only temporarily harm a small number of people, and the beneficiaries will be more, that may not be true.

　　In the first industrial revolution, the living standard of ordinary people in Britain did not improve, and even went backwards. According to the research of economic historian Charles H. feinstein, from 1770 to 1840, the wage increase of British workers not only failed to catch up with GDP, but also significantly lagged behind the price increase. Due to poor working conditions in factories and poor sanitary conditions in rapidly expanding cities, the average life expectancy of British people living in industrial cities has declined. Correspondingly, most of the wealth growth brought by technological progress has been taken away by industrialists. It was not until the completion of the British industrial revolution in 1840 that ordinary people gradually felt the dividends brought by technological changes, but the previous "short" 70 years were enough to span a person’s life.

　　Moreover, the dividends brought by technological progress will not be guaranteed to be shared equally with everyone. Blanco Milanovic, an economist who is famous for putting forward the "elephant curve", pointed out that,During the three industrial revolutions in history, income inequality increased twice.

　　Milanovic said: "In the computer age, the increase in inequality is largely due to the fact that new technologies have greatly rewarded those with higher skills. At the same time, with the middle-income routine jobs being crushed, unskilled workers have moved to low-paid service jobs, which has further widened the wage gap." Now AI is likely to affect some white-collar jobs just as the computer revolution has replaced some blue-collar jobs.

　　Although Sam Altman, the founder of OpenAI, promised that as AI brings a lot of productive resources, "it can promote human society to be more equal". But looking back at the industrial revolution, technological innovation can’t guarantee this.It may be factors other than technology that caused the Gini coefficient of Britain and the United States to drop significantly during the second industrial revolution.Karl B Frey mentioned in his book The Technical Trap that two world wars and the Great Depression destroyed the wealth of the rich, and the subsequent rise of the welfare state also promoted the distribution and adjustment of social wealth.

　　On June 20th, Huang Qifan, executive vice chairman of the Academic Committee of China National Innovation and Development Strategy Research Association, warned in an article published on the Outlook think tank that AI might trigger the fourth industrial revolution: "Without government intervention, artificial intelligence will greatly increase social inequality and bring greater class gap."

　　Technology has been reshaping our work.

　　In May of this year, Microsoft released the annual work trend index report for 2023, focusing on investigating employees’ views on artificial intelligence. The results show that 49% of the respondents are worried that AI will affect their positions.

　　This kind of anxiety is largely based on the fact that AI will lead to the disappearance of its own position. The report released by Goldman Sachs in March predicted that in the future, about two-thirds of jobs in European and American countries will realize some degree of AI automation.

　　But in fact,Technological progress has been reshaping our job requirements.For example, agriculture, which used to consume a lot of labor, has released a lot of manpower because of the emergence of machines like tractors; The development of household appliances has also liberated a large number of women from unpaid housework and enabled them to enter the workplace. Today, white-collar workers are accustomed to the office cubicle, and its invention history is less than a century.

　　Technology will not only affect the number of jobs absorbed by various industries, but also have a significant impact on various jobs themselves.

　　Before 1980s, because the technology in color negative was not popular, ordinary people had to rely on the colorist in the photo studio to get color photos. This profession is mainly responsible for coloring black and white photos. Because of the high difficulty coefficient and long operation time, the price of color photos is very expensive. However, with the advent of color photo technology and the appearance of digital cameras, color photos are no longer rare, and the post of colorist has quickly faded out of people’s memory. However, the iterative upgrade of camera technology has also led to more demands for the original monotonous studio photography, and new jobs such as post-retouchers and travel photographers have emerged.

　　Therefore, unlike people’s intuitive cognition, the new jobs brought by technological development are actually much more than the jobs it has eliminated. Economist David Otto and others found in a study last year that,At present, 60% of jobs did not exist in 1940.

　　Even for posts with the same name, with the development of technology, the tasks it undertakes are changing in different periods.

　　For example, before the advent of personal computers in the 1980s, the duties of a secretary basically revolved around typewriting and filing materials. However, with the application of computers in the office, the typing and filing jobs that once occupied a lot of secretary’s working time have decreased significantly.

　　In contrast, secretarial work has shifted from serving one person alone to assisting the whole department or organization. In 1970, 70% of secretaries worked for only one person. In 1987, a secretary supported an average of 5.3 people, but in 2001, the number rose to 12.3.

　　The picture on the left shows the secretary recruitment notice in a newspaper in Florida, USA in 1966, which frequently mentions typing skills; The picture on the right shows the secretary recruitment notice in a newspaper in Florida in 2006, asking for more skills.

　　Nowadays, with the emergence of AI tools that are good at voice transcription, meeting minutes and even coordinating work schedules, the post of secretary may change again.

　　AI’s reshaping of work will not happen overnight.

　　In 1858, an American named hamilton smith made the world’s first washing machine in Pittsburgh, which immediately caused many people’s panic and worried that women would lose a means of making a living. In a subsequent article, The New York Times comforted everyone not to worry about the impact of the new machine: "The washing machine can reduce labor and save manual work, so that we can get rid of the fatigue and many unpleasantness caused by hand washing. But the washing machine is not used to replace and will not replace the work of single young women. "

　　Facts have proved that the prediction of this article is very accurate. In fact, the number of female laundry workers in the United States has been rising for 50 years after the appearance of washing machines. The reason why this happens is that the washing effect is not good, and even the clothes will be damaged. Another important reason is that it is too expensive. Until the 1920s, when the price of the cheapest washing machine dropped to an average person’s three-week salary, this machine, which could save a lot of housework time for human beings every day, quickly appeared in ordinary families, and the number of female laundry workers also dropped significantly.

　　A similar principle applies to all kinds of new technologies.The polishing of technology, the matching of mechanism and the cost of use all make many new technologies unable to play the expected power in the workplace immediately.The technology of automatic telephone appeared in 1892, but the profession of telephone operator completely withdrew from the historical stage until after 1980, and it took nearly a century for new technology to replace labor. Nowadays, it seems that the computer that white-collar workers can’t leave at work has actually gone through decades from the emergence of personal computers to the significant impact on the workplace, and the biggest influencing factor behind it is the price.

　　Nowadays, it seems that artificial intelligence will set off a huge wave, and I am afraid I have to go the way of my predecessors. ChatGPT is the fastest growing consumer application in history, but now it is reported that it has fallen into a situation of weak growth. According to the website data analysis tool SimilarWeb, the growth rate of this chat application has dropped from 131.6% in January this year to 2.8% in May.

　　In early June, Morgan Stanley released a survey report on the use of large models, showing that the daily usage rate of both ChatGPT of OpenAI and Bard of Google is less than 5%. Many people have tried it once or twice and never used it again. "At first, I felt so smart to chat with it. I can talk like a human being, but I really want to work. I can’t do my work at present, not only because I don’t understand it well, but also because I am full of address unknown." A practitioner engaged in data analysis told The Paper.

　　Therefore, although the AI wave is coming, and it may eventually reshape our work, in most positions, this process will not be completed as soon as you finish reading for four years.Instead of obsessing about which major to choose will not be robbed by AI, what we should grasp more is to polish those abilities that AI can’t fill.

　　So what ability can’t be replaced by AI at least in the short term? By analyzing the work activities in 2017, we can conclude that in addition to the work to get started,What AI is not good at at at present is the interaction with people, which includes both daily communication and coordination and management. What AI is best at is processing a lot of information and data.

　　Therefore, even if AI has passed the accounting exam, people in this industry are still not very worried: "Accounting audit will not disappear for the time being. Accounting audit is not making decisions most of the time, but doing manual work, communicating and coordinating investigations, etc."