The World Economic Forum (WEF) released the January “Chief Economists’ Outlook” on the occasion of the Davos Forum 2024. Launched against a backdrop of prolonged weakness in global economic conditions and growing regional divergences, the report says the uncertainty that has dominated the outlook over the past year continues to cloud near-term economic developments: 56% of economists expect the global economy to weaken. next year, but 43% expect conditions to be unchanged or stronger. In this context, AI can push the boundaries of productivity. Rapid advances in AI represent another potentially profound shift for global economic activity.
The advent of generative AI, a disruptive technology
The WEF report highlights how, despite a multitude of use cases – from precision agriculture to automotive and logistics – AI adoption has remained limited in scale since the learning boom The launch of large language models (LLM) and generative AI tools, such as ChatGPT and Bard, introduced radically new possibilities at the end of 2022, opening the technology to the general public. In the May 2023 edition of the Chief Economists' Outlook, generative AI was already highlighted as a major technological shift that posed new structural challenges for businesses and policymakers, with 42% of respondents predicting the technology will become commercially disruptive by the end of 2023. Although this percentage increased to 50% for 2024 in the latest survey (see Figure 1), the results reflect an overall ambiguity about the trajectory and scale of adoption of the Generative AI by businesses, with 37% of respondents remaining unsure and a further 13%. who do not agree.
Perspectives from chief economists : Divergences between different income groups
The latest survey highlights a divergence in results between different income groups regarding the impact of generative AI on productivity over the year, with 79% expecting increased production efficiency in High-income economies respect only 38% of low-income economies (see Figure 2). The time frames within which productivity gains from AI are expected to become tangible are also very different.
Leading economists are almost unanimous (94%) predict that productivity improvements will become economically significant in high-income economies over the next five years , and 57% expect benefits to appear over the next three years. coming years. This contrasts sharply with more conservative views on the prospects for low-income economies, where only a slight majority (53%) expect benefits to be seen within the next five years and 47% believe it will take longer of five years. . No interviewee said productivity gains would never materialize, reflecting the expectation that AI will have a lasting and far-reaching impact on the global economy. One estimate suggests that generative AI alone could boost global productivity growth by 1.5 percentage points per year and increase global GDP by 7% over a 10-year horizon. . According to the most optimistic scenarios, widespread diffusion of AI could enable an increase in global production of up to 30% by the end of the century.
The geopolitics of AI
The geopolitics of AI and different approaches to regulation will define the scope and geography of AI adoption in the years to come. When asked which regions are expected to see a significant increase in productivity from greater adoption of AI, leading economists undoubtedly predict that It is the United States, China, Europe, East Asia and the Pacific will benefit the most over the next three years. . This is largely consistent with the geography of digital penetration and estimates suggesting that advanced economies are, on average, better positioned to reap the benefits of AI.
Although the definitive leaders have not yet been established, the AI scene has been largely dominated by the United States, with the country's AI-related startups attracting $26 billion in venture capital in 2023, which is six times more than those based in China. closest competitor. When it comes to the potential impact on productivity across sectors and industries, leading economists expect widespread benefits across sectors. The most significant improvements are expected in knowledge-intensive sectors, including digital information and communications technologies, financial and professional services, medical and health services, retail, manufacturing , engineering and construction, energy and logistics. The views of leading economists are reflected in industry estimates, which suggest a possible 5% increase in annual turnover in the banking and pharmaceutical sectors.
Nearly three-quarters of the productivity gains enabled by AI across all industry sectors will be enabled by improvements in research and development, customer services, marketing and sales, and software engineering. Positive predictions about the economic gains of AI have been accompanied by widespread concern about the technology's possible implications for society, employment and inequality. Although estimates could quickly be outpaced by the pace of evolution in LLM sophistication and the scale of AI adoption, the latest survey of leading economists assesses the scale of the impact possible from recent developments in generative AI. The results reveal a notable divergence in outcomes between high- and low-income economies (see Figure 3), consistent with different prospects for productivity gains.
Nearly three-quarters of respondents expect generative AI to accelerate the pace of innovation (74%) in high-income economies, while only 31% expect a similar effect in economies low income. While innovation and productivity gains through AI attract greater investment, the slower pace of progress in lower-income economies risks widening the technological and economic gap. The gap is also visible when looking at living standards: 57% expect AI to increase in high-income economies, compared to 41% in low-income economies.
Chief Economists' Outlook : Concerns About Jobs Risks
The potential gains from productivity gains contrast sharply with concerns about the risks of automation, offshoring and job degradation. According to the Chief Economists' Outlook, Nearly three-quarters (73%) of leading economists surveyed do not expect a net positive impact on employment in low-income economies. , while a further 17% are unsure. In other words, the majority expects change. Views on high-income economies are slightly more mixed, with 47% thinking net job creation is unlikely this year and 30% unsure. . One estimate suggests that AI could automate up to 18% of work globally, with advanced economies and occupations primarily involving routine, repetitive tasks having been most exposed so far. Even as evidence of the broader effect on the workforce continues to evolve, there are signs of consensus that AI will likely transform rather than destroy jobs in the short term with a potentially negative impact on quality of work.
Another area of uncertainty concerns the prospects of generative AI leading to a significant relocation of economic activity or a decoupling of productivity and wages. Chief economists are roughly evenly distributed across high-income economies. For low-income economies, a slight majority said the reshoring of economic activity (58%) and the decoupling of productivity from wages (54%) are unlikely. Recent data suggests that workers exposed to AI have thus far enjoyed higher wage premiums.
The long-term impact on wages at different skill levels will depend on the extent of productivity gains and changes in the quality of work. Perhaps the most immediate and acute impact will be felt in the public and political space, with one estimate suggesting that up to 90% of online content could be generated by AI by 2026.
The delay in regulating AI
Although AI regulation lags behind adoption, generative AI tools have been found to replicate historical patterns of inequality and social bias. Concerns about the risks of misinformation have also increased in the run-up to elections around the world next year, according to the Chief Economists' Outlook . For example, the number of websites publishing AI-created misinformation has already increased by more than 1,000% since May 2023. The survey results reveal this concern, with 56% and 44% saying that the Generative AI will lead to a significant decline in trust in high-income and low-income economies respectively this year.
The lack of consensus on some of the perspectives discussed above reflects ambiguity about the trajectory and scope of AI's impact on workers and economies. Decisions made today will change this trajectory and their impact will be felt on individuals, societies and economies for years to come.
Comments
Post a Comment