Fanny Wu explores how AI affects future job structure, social inequality, and how we can better equip ourselves to embrace the ongoing challenges.
Do people think their hard-earned job could be replaced by the work of a machine? The status of having a job would no longer be a major distinguishing factor between people; rather, the workforce would be separated into two major groups, highly paid, skilled workers and low-paid, unskilled workers, by their routine-or-not characteristics. This phenomenon is widely discussed as ‘job polarisation’.
While AI could take jobs from middle-skilled workers, the extra wealth generated could incentivize consumer demand, increase consumption, and create jobs elsewhere. Transitions in the labour market will also occur to meet the new demand in other areas. Thus, it’s easy to understand why optimists claim there will not be a serious unemployment issue in the long run. The only problem is that this conclusion is only valid if the speed of jobs being generated is at least as fast as that of them being taken away. If not, it would increase the unemployment rate and create a discouraged worker effect, which is a bigger problem in the long-run since it reduces the labour force participation rate, and could be an obstacle to healthy economic growth.
Moreover, data and common-sense show that with deep learning — a technique that allows systems to learn and improve by analysing data and examples rather than being explicitly programmed — AI can evolve more quickly than humans, who need time to learn and improve. Therefore, AI could replace a greater proportion of the workforce in a shorter time, and lower the value of human labor in many sectors. This could discourage young people from developing skills that they haven’t already acquired, which will be especially problematic for those from lower income backgrounds. As people with lower income are more risk-averse, have lower budget constraints, and face a more elastic part of the indifference curve between income and leisure, they are very sensitive to uncertainty and the common belief of future income. Rapidly changing technology makes it harder for people to predict which skills they may need or to understand the prospective labour market. Such uncertainties about future industrial practices, make it seem rational to maximise utility by reducing the cost of time, money, and effort in higher education.
The economy could potentially diverge to a point where only a minority of exceptionally talented people earn super high wages while others only earn a pittance in comparison. Income inequality not only matters for the growth, sustainability and economic stability; it also negatively affects productivity by inefficient capital-labor matching. Furthermore, it can concentrate political and decision making power in the hands of a minority, causing a biased understanding of people’s need and subsequently unrepresentative governance.
Cross-country inequality is another potential issue. Most of the money made from artificial intelligence will go to the United States, China, and other nations with successful AI businesses that are conducting further research and investment. These countries are already using AI and expanding operations to other countries, and could potentially monopolize the AI market. At the same time, globalisation has expanded the scale of these winner-takes-all markets, enabling vast salaries and profits to be shared among a narrow set of employees and shareholders such as Google, Amazon, and Alibaba. By using these technologies, citizens of these nations can move between jobs and gain skills more easily. They can also take greater advantage of increased finances — earned by efficient adaption of new technology — to further increase their income and earnings potential. On the other hand, those nations without this technological background must sacrifice trade liberty to negotiate and essentially determine their choice of geopolitical alliances by becoming economically dependent on their AI supplier—potentially altering the global balance of power.
The development of AI technology is exciting and scary. While understanding the technological development is important, we must also consider potential gender parity issues in a workplace dominated by the STEM industry, and immigration policy in a technologically stratified world among other important sociological topics, to help us better estimate and prepare for the future. Yet, no matter how the workplace may change, having a solid foundation of basic literacy and numeracy skills and other ‘character skills’, such as perseverance, sociability, and curiosity will be crucial for us to switch from routine, unskilled jobs to non-routine, skilled jobs. Doubtlessly, an upgraded educational system and flexible government policy to make it easier for workers to acquire new skills and switch roles, are necessary for long-term development.