Where is the AI Supercycle Headed?
Given the current pace of advances in generative artificial intelligence, it is difficult to predict how the technology will impact the economy, business, and society, but it is already clear that new AI applications will create a few winners, diminish the workforce, and pose significant policy challenges for governments.
Consider how AI impacts the three main drivers of growth: capital, labor, and productivity. On the capital side, the massive investments required to drive AI innovation will ensure that there will be a small, concentrated set of winners. Large technology companies with monopolies in their respective markets are the only ones that can afford the enormous costs associated with developing, training, and promoting language models at scale.
Most of these costs come from running high-performance graphics processing units and powering and cooling huge data centers. Multiverse Chief Technology Officer Sam Mugel estimates that training the next generation of large-scale language models will soon cost at least $1 billion. In 2023 alone, the so-called “Magnetic Seven” (the top US technology companies) have allocated a combined $370 billion to research and development, roughly the same as the EU's total R&D budget (counting both corporate and public sector).
When it comes to labor, it is too early to predict winners and losers, or how AI-related gains and losses will be distributed across the economy. A 2023 Goldman Sachs report estimates that AI “could lead to the automation of 300 million full-time jobs,” but a World Economic Forum survey of 803 companies noted that job creation associated with investments in the green transition and climate change adaptation would result in much lower net losses.
Either way, many are concerned that AI will contribute to long-term structural unemployment, creating job losses for both skilled and unskilled workers. But while the above predictions provide a baseline of what could happen, there is plenty of room to further refine our thinking on this issue. Ultimately, the scale of the problem will depend on which jobs are lost at various points in the AI value chain.
It remains to be seen how job losses in one link in the chain will impact the rest of the technology sector, much less the economy as a whole. The impact on jobs is likely to vary significantly as it moves from chip manufacturers, AI infrastructure, and AI applications to sectors such as healthcare, education, and communications, all of which are poised to benefit from AI innovation. At the base of technology, we are already seeing significant growth and job creation as chip manufacturers (such as Nvidia) build manufacturing facilities and invest in the production capacity that will drive the AI revolution.
How many jobs will be created or lost elsewhere is less clear, since we can’t predict all of the ways new technologies will be used and what knock-on effects they will have. Early signs of AI’s impact on long-term efficiency and productivity gains are encouraging, at least for workers who still have jobs. For example, a 2023 study of 5,000 workers by Erik Brynjolfsson, Daniel Lee, and Lindsay R. Raymond found that AI tools increased worker productivity by 14% on average, and 34% for new and low-skilled workers.
Many fear that long-term structural unemployment will create a pool of unemployed workers, both skilled and unskilled.
Dambisa Moyo
Technological advances have a long track record of strengthening global connections in trade and communications, expanding access to public goods like health care and education, driving innovation, improving living standards, and ultimately stimulating broad-based economic growth. There is no reason to think that AI won't do the same.
Moreover, AI is likely to diffuse across the economy faster than previous technologies, so AI-related productivity and efficiency gains may come sooner rather than later. Previous general-purpose technologies (such as the steam engine, electrification, and personal computers) required significant expenditures to build the underlying infrastructure. It took more than 40 years for electricity to become widely available in the first half of the 20th century, and about a decade for smartphone penetration to exceed 90% in the 2010s. In contrast, AI can be deployed through existing digital platforms and devices.
Ultimately, the AI supercycle is likely to drive increased productivity and enhanced economic growth, which PwC projects could reach $16 trillion globally by 2030. However, these gains will accrue primarily to capitalists and not to a potentially shrinking workforce. In an era of labor-intensive growth, many companies and industries will need to adapt their business models, specifically increasing the capital-to-employment ratio, forcing governments to reevaluate tax and welfare policies.
If greater economic benefits flow to capitalists, tax systems will need to change accordingly. For example, corporate tax rates may need to be significantly increased to capture the excess profits generated by automation and the shrinking workforce. On welfare, the threat of increased structural unemployment due to AI will reinvigorate debate around previously radical proposals such as universal basic income.
We need to reflect on the impact of AI on inequality both within (capital and labor) and between countries. A widening gap between tech leaders such as the United States and China and the rest of the world (especially the poorest countries) bodes ill for an already tense geopolitical environment.
• Dambisa Moyo is an international economist and author of four New York Times bestselling books.
©Project Syndicate
Disclaimer: The views expressed by the authors in this section are their own and do not necessarily reflect the views of Arab News.
