The future of work is at stake as artificial intelligence (AI) continues to advance, and the implications are profound. The launch of ChatGPT in 2022 sparked a wave of excitement and concern, with AI bosses warning of an impending jobs apocalypse. While the labor market in the rich world is currently strong, with record-high employment, the fear is real. Seven in ten Americans believe AI will make finding work more challenging, and nearly a third worry about their own job security. The situation is exacerbated by a shortage of job openings for college graduates, particularly in computer programming.
However, history provides some perspective. Technological progress has never significantly reduced the overall demand for human labor. The Industrial Revolution, despite its impact on working-class wages, did not diminish the need for human workers. Yet, the current AI boom is different. The top AI models are incredibly advanced, capable of handling complex coding tasks that were unimaginable just a year ago. The number of AI agents is exploding, and business spending on AI is skyrocketing. Annual recurring revenue for Anthropic, a prominent AI model developer, is projected to reach a staggering $50 billion by the end of June.
Despite the lack of evidence in labor market data, the rapid advancement of AI raises concerns. Society may be on the cusp of a significant reallocation of resources and political upheaval. Economists' predictions that work will remain plentiful are less reassuring when considering the long-term implications. As AI models and robots become more capable, the quality and wages of human jobs are not guaranteed. Data centers, for instance, will consume a substantial portion of America's peak power demand in 2027, according to Goldman Sachs. As AI firms drive up land and energy costs, the purchasing power of human earnings will diminish.
This dystopian scenario is prompting discussions about state intervention and universal basic income in Silicon Valley. While such measures are far from being implemented, governments may need to act sooner than expected. The 'China shock' of the late 1990s and early 2000s, which led to the loss of approximately 2 million American jobs due to China's entry into global trade, serves as a cautionary tale. Despite the relatively small number of job losses, it fueled public anger and contributed to the election of Donald Trump, who imposed the highest tariffs since the 1930s.
The white-collar employees threatened by AI possess more political and social influence than factory workers affected by Chinese competition. Even a small number of job losses could trigger a backlash against AI technology. The opposition to new data centers is a glimpse of the potential unrest that may follow. Severe disruptions to the security and status of many people could lead to widespread civil unrest, even revolution.
Governments have several options to address these challenges. One approach is to slow down the pace of change, but inhibiting technology is not advisable. AI offers immense benefits, including increased wealth and progress in combating diseases and addressing global issues like climate change and poverty. The Luddites' attempt to halt the automation of textile mills in 19th-century England had dire consequences for the world.
A better strategy involves implementing clever tax reforms. If employment declines, the income that once went to workers may be captured as high profits by AI firms, chipmakers, data centers, or other supply chain entities. Levies on corporate profits above a normal return on capital, land, and natural resources could help capture these rents. Inheritance taxes to prevent the concentration of wealth among the elite also become more relevant.
Additionally, governments can assist workers in adjusting to the changing landscape. Public wage insurance, which provides income support during job transitions, can help workers find better opportunities. Denmark's active labor market policies, which assist individuals in finding and training for new occupations, have proven effective in reducing unemployment durations.
These measures would enhance economic efficiency and fairness, regardless of AI's impact. However, in an era of populism, technocratic reforms may struggle to gain public support. Past attempts to help workers adjust to trade liberalization have met with limited success. In a future dominated by AI, humans will require assistance in surviving, not just adjusting.
Radical ideas, such as the partial nationalization of AI firms, have been proposed. While a South Korean presidential adviser's suggestion of a citizens' dividend from AI businesses caused a stock market downturn, similar concepts are discussed in the United States. In economic terms, a well-designed tax system and government ownership in the private sector can achieve similar goals. However, America may find that some public ownership is essential to ensure the social benefits of AI are transparent.
The concentration of rent must be addressed early to prevent the power of rentiers from becoming too entrenched. The jobs apocalypse is not yet upon us, but governments should not wait for conclusive evidence before creating a safety net. It is better to start preparing now to ensure a smoother transition into the AI-driven future.
