Programmer, “Slow Death”

“AI has not changed the business model of the film and television industry. It has only squeezed out the existing market share by compressing labor costs to the extreme.” Data from the DataEye Research Institute shows that in the first half of 2026, there were more than 221,900 new AI short dramas launched on the entire network, but only 1,055 of them exceeded 100 million views, and the rate of exceeding 100 million was only 0.47%. More than 90% of small and medium-sized producers did not make any money.

Hong Ping said that AI drama companies will either use the “drama sea tactic” to bet on one or two to become a hit, or they will continue to reduce costs and increase efficiency, lowering manpower and computing power costs. These are all unsustainable stocks.

Also feeling the AI ​​crisis is the programmer industry. For a long time, “switching” has been a career change choice for a large number of science and engineering students and a way to earn high salaries, attracting young people who are not ideal in school or majors to enter this field to earn more money while they have the physical strength. Data from the Ministry of Industry and Information Technology shows that as of 2024, the number of developers has exceeded 9.4 million, with an average monthly salary of 15,302 yuan, and salary levels have been at the forefront of the industry for many years.

However, the advent of AI programming has directly lowered the learning threshold for programming. Many corporate positions, such as front-end, back-end, testing, operation and maintenance, etc., have begun to be merged and abolished. Liepin Big Data Research Institute’s “Programmer Transformation Trend Report in the AI ​​Era” (2026) shows that from January to September 2026, new positions for traditional programmer positions dropped by 8.0% year-on-year.

“With the help of AI, one person can do the work of two or three people, and companies use previous job descriptions less and less in recruitment.” Yu Yang, CEO of Chuanzhi Education, told China News Weekly that although AI has impacted traditional programming positions, it has spawned many new positions, such as AI full-stack engineers and FDE engineers for large model development, and embodied intelligent physics AI development.

Yu Yang revealed that the salaries of these new positions are very competitive. In addition to traditional IT Internet practitioners and college students majoring in computer science, 40% to 50% of students who sign up for AI transformation training are cross-industry personnel. Everyone’s goal of learning AI skills is very clear, which is to increase salary, avoid elimination, and keep up with the market trend of “the Internet is going down and AI is going up.”

However, in the face of rapidly iterative AI technology, many people in the industry are worried that newly learned skills will be discarded as soon as they are learned. The reporter learned from Github, the world’s largest programmer community, that technical discussions in the field of large model application development and Agent workflows continue to be heated, but the focus of discussions frequently shifts, from model fine-tuning, to RAG, to Agent orchestration and tool invocation, and almost no technical route can maintain its popularity for more than half a year.

“Everyone knows that AI is the general direction, but everyone is still waiting to see which direction to choose. Our AI courses will be updated four times a year in order to adapt to the latest job requirements in the industry.” Yu Yang said that when the big model first came out last year, it was easy to find a job by optimizing the model side, or simply doing RAG and intelligent agents. But this year, many companies have integrated large models with products, and the requirements for large model application development have become higher. In addition to understanding large models, they also need to understand specific products and scenarios.

“AI Whitewashing” Phenomenon

“So far, there are no signs of obvious employment impacts caused by artificial intelligence from the macro data, but there are already signs at the micro level.” Cai Fang pointed out.

A special survey of 8,000 employers in 14 countries around the world by the Organization for Economic Cooperation and Development (OECD) showed that 11% of companies that have adopted AI have reduced employment, of which only 9% attributed layoffs to AI. Roughly speaking, among all enterprises that adopt AI, only about 1% of employees will be reduced due to AI.

“This is a short quiet period.” Cai Fang believes that AI iteration is too fast and future expectations are unclear. Companies generally choose to wait and see: reducing external recruitment and not taking the initiative to lay off employees for the time being. However, it is not clear which direction to train existing employees, forming a stalemate.

In the white-collar industry, this job shrinkage caused by AI replacement is particularly significant.

“People may not be laid off in large numbers, but the recruitment of traditional positions is decreasing.” Jiang Kaifeng, a professor at the Department of Organization and Strategic Management at Peking University Guanghua School of Management, told China News Weekly that AI is changing companies’ needs for middle-level management positions. Many tasks such as communication and coordination, information transmission, and process monitoring can be partially completed by AI. AI is also promoting the integration of technical positions. For example, some companies have begun to merge several previously scattered positions such as software and hardware sales, installation, debugging, and maintenance into one “problem solution engineer” or “problem solution sales” position.

Jiang Kaifeng also observed that although the names of some positions have not changed, the actual work content and skill requirements have changed greatly. For example, in the recruitment of some financial transaction positions, the relative importance of traditional financial and accounting skills has declined, while programming, IT technology and data analysis skills are increasingly valued. “A CFO told me that their company has started to adjust its recruitment ideas for financial positions this year. It no longer only focuses on accounting majors, but is more inclined to recruit talents with computer and data analysis backgrounds. They hope to transform the professional knowledge, business logic and decision-making experience accumulated by financial personnel over the years into a knowledge system that AI can understand and use, and gradually participate in complex analysis and decision-making that in the past required senior financial personnel to complete.”

The “AI replacement” of blue-collar industries occurs at a different pace. Compared with the slow erosion of white-collar jobs as “recruitment entrance narrows”, labor-intensive manufacturing industries tend to occur intensively in the form of production line transformation and job mergers.

“AI is developing too fast, and most companies still don’t know how to integrate and embrace AI. After a large number of aimless attempts, they have not brought substantial returns, resulting in a lot of waste of investment.” Yu Yang believes.

Jiang Kaifeng observed foreign cases and found that the layoffs of some companies may be more due to operating pressure, organizational adjustments or human resource planning, but are classified into the category of AI transformation. “If companies overemphasize the role of AI in layoffs, or even package general business adjustments as the results of AI transformation, there may be a phenomenon called ‘AI-washing’.”

“Our research has found that although news of ‘AI layoffs’ have emerged one after another in recent years, most of them are concentrated in a few large companies. In contrast, many small and medium-sized enterprises have not really started to apply AI, or although they have started to try, they are still in the stage of scattered use and lack of systematic planning. Some companies believe that they only have two or three hundred people, and there is no need to invest a lot of money in AI transformation.” Jiang Kaifeng suggested that to have a more comprehensive understanding of the impact of AI on employment, the data of small and medium-sized enterprises may be more representative, and they cannot just focus on large companies.

In response to the domestic situation, Jiang Kaifeng’s team used the “AI Occupational Exposure Index (AIOE)” to analyze the degree of overlap between the capabilities required for recruitment positions and AI capabilities based on approximately 1.45 million job advertisements from domestic listed companies from 2016 to 2025. The study found that the average AI exposure index of recruitment positions in the sample peaked in 2019 and has generally declined since then. In terms of industries, the financial industry has the highest index, followed by education and information transmission, software and information technology services, and the transportation, warehousing and postal industry, and accommodation and catering industries have lower indexes.

“This indicator reflects the degree of overlap between the capabilities required for the job and the AI ​​capability, not the proportion of positions being replaced.” Jiang Kaifeng explained that the decline in the index may mean that companies are reducing recruitment for some jobs that originally needed to be done by humans but are now increasingly capable of being performed by AI.

How to implement the “unemployment buffer period”

On October 23, the General Office of the State Council issued the “Opinions on Adapting to the Development of Artificial Intelligence Technology to Help Build an Employment-Friendly Development Model”, proposing for the first time that “enterprises are encouraged to explore the establishment of an unemployment buffer period system, inform affected workers in advance during the large-scale promotion and application of artificial intelligence, and provide support for transition and employment.” This is the first time that the “unemployment buffer period” has appeared in a national-level policy document.

“The impact of AI is in a brief quiet period, which gives us a rare opportunity to prepare countermeasures in advance.” Cai Fang said.

Existing judicial and local practices have begun relevant explorations. The case of Zhou in Hangzhou is currently a precedent with signaling significance, and is known as the first “AI replacement case” in Hangzhou.

35-year-old Zhou was originally the quality inspection supervisor of AI large models at a financial technology company in Hangzhou. In January 2025, the company proposed to transfer him from a supervisor to a general operation position, with his monthly salary reduced from 25,000 yuan to 15,000 yuan. After the negotiation failed, the company unilaterally terminated the labor contract on the grounds that he was “impacted by AI technology.” From the labor arbitration to the first and second instance, it was determined that the company had illegally terminated the company and was awarded compensation of more than 260,000 yuan. The Hangzhou Intermediate People’s Court made it clear in the second instance that the introduction of AI to realize job replacement and complete technology upgrades does not constitute a “major change in objective circumstances” stipulated in the Labor Contract Law, and the company cannot unilaterally terminate the labor contract on this basis.

Jiang Xiaotong, the attorney representing Zhou in the case and a lawyer at Yufeng Law Firm, told the China News Weekly reporter that in this case, the salary decrease was a very core fact for the court to review the rationality of the job transfer. Many employers will not write “AI replacement” into the termination notice at the beginning, but wait until the arbitration and litigation stages to take it out and package it as objective business reasons. The essence is to use new technology as a reason to reduce the cost of illegal termination.

“The use of AI belongs to the enterprise’s operational autonomy, but it does not mean that it can be used to fire employees at will. Now the court will not accept it just by hearing the words ‘AI will replace the job, and the job will be gone.'” Jiang Xiaotong said that whether it can be established depends on whether there is a real deployment, whether the position is actually eliminated, and whether there is a negotiation for job relocation first. AI cannot be used as a simple excuse for exemption.

Jiang Xiaotong further introduced that in this type of case, it is most difficult for workers to prove two things: first, that the company’s so-called “AI replacement of jobs” really happened, and not just as an excuse to reduce costs; second, whether the company maliciously transferred jobs and reduced wages in the name of AI to force employees to leave. She suggested that two points can be reflected in the “buffer period”: first, the obligation to notify in writing in advance, specifying which positions are to be optimized with AI, the scope, and time; second, the job transfer and training plans should be made into written materials that can leave traces and can be verified to facilitate subsequent evidence and reduce obstacles for workers to obtain evidence.

At the local level, explorations similar to the “buffer period” of AI have been carried out in labor-intensive areas. Jiangsu Province is following the path of “setting a bottom line for industry operations”. Su Zhengbanfa [2026] No. 48 is the first document in the country to “explore the reasonable setting of minimum standards for labor participation rates in labor-intensive industries and the establishment of an unemployment buffer period system”. It also clarifies that “establish a special investigation system for the impact of artificial intelligence on employment, and continue to track the penetration of artificial intelligence applications, job creation and replacement, changes in skill demand, etc.”

“When and under what circumstances the buffer period needs to be initiated, multiple aspects of information can be considered comprehensively, rather than solely based on the company’s active declaration.” Jiang Kaifeng suggested that the conditions of companies in different industries and different sizes vary greatly, and unified standards cannot be simply adopted, nor can it be assumed that just because companies have increased investment in AI, it will definitely reduce employment.

The exploration of leading companies in dealing with AI has certain demonstrative significance. IBM provides a reference similar to a “buffer period”. According to Jiang Kaifeng, IBM plans to triple the recruitment of entry-level positions in the United States in 2026 and redesign the work content of these positions to allow young employees to reduce repetitive work and participate more in customer communication, problem solving and other tasks that require human judgment and creativity.

Jiang Kaifeng believes that this is actually a choice worth considering. “AI has been able to complete many tasks that used to be undertaken by junior employees. It seems that companies can hire fewer people in the short term. But if we don’t train young people today, where will we get management talents who are familiar with the company’s business and culture in three to five years?”

(Hong Ping is a pseudonym in the article)

Published in the 1259th issue of “China News Weekly” magazine on November 9, 2026

雜志Title: AI replaces employment, the country takes action

Reporter: Li Siyang

Related Reading

• World, line up to sell goods to China!2026-11-12
• Ni Runfeng, the “King of Color TV”, passed away. From a workshop technician to the head of a state-owned enterprise, his ups and downs in half a life defined an era of home appliances.2026-11-12
• Original “Fearless” Wang Jianlin regains control of Wanda Business Management, and his net worth has evaporated by 200 billion in ten years2026-11-12
• Original: The middle-class people are crazy about grabbing Archaeopteryx in Sam2026-11-12
• There is no more double 11 in the world, leaving businesses miserable2026-11-11