The 2024 World Summit on the Information Society (WSIS) and AI for Good Summit have provided a comprehensive view of the economic impact of artificial intelligence (AI). While the potential benefits of AI in driving economic growth and productivity excite investors and stock markets, the summits also highlighted significant challenges and potential negative consequences. AI has potential for productivity growth but could also amplify the next economic downturn, by disrupting particularly labor markets.
Driving Productivity and Economic Growth: A Boost with Caveats
AI's impact on productivity and economic growth is transformative. AI technologies are streamlining operations, reducing costs, opening new revenue streams and even creating new markets and industries. In the healthcare sector, for instance, AI-driven diagnostic tools are reducing diagnosis times and improving accuracy, thus enhancing patient outcomes and operational efficiency. Similarly, in finance, AI algorithms are streamlining trading operations, fraud detection and risk management processes. Andrew McAfee argues that generative AI can be considered a general-purpose technology like the steam engine and electrification, fueling the next wave of economic growth.
However, these productivity gains come with significant human costs. Up to 40% of all working hours could be impacted by large language models (LLMs), as reported by the World Economic Forum. Automation of clerical tasks is already leading to job displacement in many sectors, from retail to banking to the software industry. While new jobs are also being created, the speed at which AI is impacting labor markets – which are typically slow to adapt to technological changes – may exacerbate the next economic crisis.
Automation, Downturns and Job Displacement
Research shows that 88% of automation related job losses occur in the first year of recessions. After the Global Financial Crisis, rather than rehiring workers after the slump, many firms automated their operations. This has led to the most severe "jobless recovery” ever seen in the US and Western Europe, driven almost entirely by the loss of routine jobs. Today, AI is likely to threaten a wider range of jobs than in the past and those more at risk of displacement are those with repetitive cognitive tasks, which are more common in advanced economies.
As estimated by the IMF, jobs in advanced economies have a higher overall exposure to AI (60% of total employment) than in emerging markets (40%), which includes also those benefiting from AI. From these, the share of jobs at risk of AI substitution is 30% in advanced economies, 20% in emerging markets and 18% in low-income countries, as highlighted by Gita Gopinath at the AI for Good Summit. Therefore, workers with clerical jobs in advanced economies will be those more likely to suffer from the AI deployment.
Addressing Inequality: Potential and Pitfalls
AI could be then seen as a great leveler. By impacting proportionally more white-collar than blue-collar jobs, this wave of automation may reduce the inequality among workers with different skill levels, a distance which has been increasing over the past century. Moreover, as outlined above, it may negatively affect advanced economies more than emerging ones. The latter could also embrace (generative) AI, which promises to democratize access to information and services, to “leapfrog” into a higher stage of economic development.
Yet, the deployment of AI in addressing inequality is fraught with challenges. Emerging markets and low-income countries are less prepared than advanced economies for rapid technological changes. Almost a quarter of young people in these countries are not in employment, education or training, which makes them even more vulnerable to this shift. Moreover, the economic growth promised by AI will not be evenly distributed. While large (tech) conglomerates will take the lion’s share, smaller businesses, especially in developing countries, often lack the resources to invest in sophisticated technologies, exacerbating existing economic disparities. The risk of an increasing digital divide is becoming even more pronounced as AI adoption accelerates.
Policy and Regulation: Striking the Balance
The rapid pace of AI innovation is thus creating regulatory challenges for governments. A key takeaway from the 2024 WSIS and AI for Good Summit is the necessity for a balanced approach: policymakers must navigate the fine line between encouraging innovation and protecting societal interests:
- Review current tax systems to remove existing biases in favor automation over people. The corporate tax incentives in some countries are encouraging labor substituting investments in technology. This is not only dangerous but also inefficient from a social return perspective.
- Enhance both regulation and supervision by financial regulators, e.g. on AI models’ sources and limits, to mitigate the threat of an AI-amplified crisis. This will require themselves to upskill to better gauge AI-related risks, while AI tools can contribute to improve monitoring, risk assessments and early detection of vulnerabilities.
- Embrace AI tools to improve public administration functions, also to finance the above investments. From increasing tax compliance and widening the tax base, to better targeting social assistance, governments should be among the first adopters of AI solutions to be employed always together with human oversight.
Finally, countries that prioritize investments in education, digital competencies and digital infrastructure, are most likely to stay ahead. This applies notably to developing countries. There is a critical mismatch between the skills of displaced workers and the requirements of new AI-centric roles: reskilling and upskilling initiatives, also through AI tools, can help tackling this pressing gap.