Opinion | Demographic Economics: How does AI affect job market?
By James Liang, executive chairman, former CEO, and co-founder of Trip.com Group
The release of reports by companies and the launch of AI avatars for data interpretation exemplify the application of artificial intelligence (AI) in the workplace. With the help of AI, people can have more time to think and create. In this article, I would like to discuss the differences between artificial intelligence and the human brain, as well as how this affects various industries, particularly the job market.
Before addressing the question of AI and employment, we need to understand the similarities and differences between artificial intelligence and human brains. Both share a surprising and crucial commonality: unpredictability and difficulty in explanation. Unlike machines that operate on fixed rules and deterministic behaviors, human behavior follows broad patterns while retaining a certain degree of unpredictability at the individual level.
Traditional computer programs are predictable, but large models are not; they may provide slightly different answers to the same question. Similarly, AI algorithms are challenging to interpret. Deep neural networks consist of billions of parameters distributed across multiple layers. Each layer processes data non-linearly, making it extremely difficult to trace how a specific input leads to a particular output. The complexity of these models makes understanding their internal workings nearly impossible.
Additionally, there are key differences between AI systems and the human brain. The energy efficiency of the human brain far surpasses that of AI. The brain operates at about 20 watts, similar to a small light bulb, while large language models may require immense computing power and energy, potentially consuming thousands of times more energy for similar tasks.
Moreover, AI now lacks emotions, ethics, or consciousness—though one might wonder if it could develop these characteristics over time. Emotions arise from brain functions; theoretically, as AI advances, they can mimic emotions. However, emotions are also the result of billions of years of evolution. How long would it take for AI to develop true emotions? Furthermore, the question of whether AI can possess consciousness or ethics is a significant philosophical issue that warrants further exploration.
Lastly, although AI may surpass the human brain in certain cognitive tasks, AI-driven robots still significantly lag behind humans in dexterity and agility. For instance, while AI can perform entry-level programming tasks, it struggles with jobs requiring service, such as those of a waiter.
The complexity of human anatomy, coupled with advanced sensory feedback and exceptional motor control, gives human hands unparalleled precision and diverse movement capabilities. We currently lack materials that can functionally match the capabilities of the human hand, unlike the compatibility between chips and neurons.
Therefore, enhancing the dexterity and agility of robots presents a more challenging problem than language processing. Even the most advanced AI robots still cannot match humans in performing simple tasks, like picking up objects easily, let alone more complex tasks like cleaning a room. While robotic advancements will improve over time, this progress will be gradual, and many service jobs will remain reliant on human workers for many years to come.
Generative AI Creates More Jobs!
The varying capabilities of artificial intelligence across different tasks impact various jobs differently. Recent studies have revealed the potential effects of AI on employment, addressing widespread concerns about job loss. McKinsey's analysis predicts that by 2030, generative AI could automate 29.5% of working hours in the U.S. economy, up from 21.5% currently. Many research conclusions suggest that the impact of AI on employment will manifest in job transformation rather than outright replacement, with many positions being enhanced rather than completely disappearing.
AI's influence on the job market varies significantly across sectors. Certain jobs, such as translation, may be almost entirely replaced by AI, while most professions may experience partial automation, with specific tasks taken over by machines. However, due to inherent risks and the need for human judgment, some roles will still be performed by humans.
Most research focused on AI's impact on employment tends to emphasize automation, overlooking the crucial point that accurately predicting employment market trends requires considering supply and demand factors simultaneously. In the age of AI, some industries may see increased demand, offsetting the effects of automation. For example, even if AI automates 50% of tasks in a certain position if that industry expands threefold, employment numbers could still rise. Additionally, as society becomes wealthier and personal leisure time increases, consumer behavior is shifting towards a pursuit of experiential and spiritual fulfillment rather than material goods, a broader industry trend that may create new job opportunities.
Industries Most Susceptible to Automation
From an economic perspective, the pressing question is which jobs AI will replace and whether large-scale unemployment will occur. Which industries will be positively or negatively affected? How will AI influence innovation and education? Will AI affect income distribution?
To address these questions, I created a 2x2 quadrant chart (see the diagram), where the horizontal axis represents the degree of technological automation in industries.
Industries that are easier to automate include agriculture, home appliances, clothing, automobiles, and digital entertainment. Those that are more challenging to automate include real estate, as construction and renovation workers are hard to replace with robots in the short term. Travel services also involve human transport and service, making them difficult to automate.
The vertical axis reflects levels of demand. Human material needs encompass industries related to food, clothing, and shelter. Once material needs reach a certain saturation point, they stabilize, while spiritual needs are nearly limitless. Industries such as travel, entertainment, and education fall under spiritual needs, as innovation satisfies the human instinct for exploration.
From this, we can conclude that some industries will shrink, such as manufacturing; some will stabilize, like real estate and digital entertainment; and others, such as travel and education, may expand and absorb more labor.
Thus, the structure of employment in the overall economy will change, but it will not lead to mass unemployment. Some industries, like manufacturing, will contract, while others, like real estate and digital entertainment, will remain stable. At the same time, industries like travel and education may grow and attract more labor.
It is worth considering that with advancements in AI technology and economic restructuring, entirely new job categories may emerge. One increasingly prominent position may be that of a social media influencer, who provides recommendations in areas such as travel, food, gaming, and various experiential consumption. These individuals can generate income through advertising and sponsorships.
Overall, as people accumulate more leisure time and wealth, the future job market may encompass a range of careers that satisfy our growing desire for unique, enriching, and meaningful experiences. While AI may automate certain aspects of these roles, the human touch in creating and curating these experiences—a vital form of innovation—may still be irreplaceable, and these limitless imaginative industries will continue to generate job opportunities.
Innovation Field? Not Yet
From a long-term perspective, perhaps in a generation, even construction workers and tour guides could be completely replaced by robots and AI. However, one role may always remain human-dominated: innovation. In the quadrant chart, "innovation" is placed in the upper right quadrant because it provides the highest satisfaction and is particularly difficult to automate. Innovation is not limited to a single industry but is intertwined across various research and development activities.
As more jobs become related to innovation, AI and robots will only play a supportive role. Some argue that innovation can be achieved by a few geniuses rather than a large population; however, I believe this perspective clearly contradicts historical trends. Humanity is investing increasingly more capital and human resources in innovation, and densely populated cities and regions tend to be more innovative—this trend shows no signs of slowing down. In the future, more individuals will have the ability and willingness to participate in some form of innovative activity, including both high-skill jobs (like AI programming) and low-skill jobs (like game testing and movie reviewing).
Moreover, AI requires vast amounts of user-generated data for training, and a large consumer base is a crucial factor for innovation in the AI era.
The advancements in technologies like AI mean that humans can create more goods and services with less working time, which is a boon for humanity. If AI makes humans or the population redundant, it indicates a problem with the work and social distribution systems that need improvement, focusing on how to better allocate goods and services to everyone.
Related News:
US Department of Justice urges Google to sell Chrome in antitrust move
Comment