- By Annabelle Liang
- Business Correspondent
Kristalina Georgieva, managing director of the IMF, says, “In most circumstances, AI will worsen overall inequality”.
Ms Georgieva adds that policymakers need to address a “complex trend” to “prevent technology from fueling further social tensions”.
The proliferation of AI has brought into focus its benefits and risks.
The IMF said AI will impact the most – 60% of jobs in advanced economies. In half of these cases, workers can expect to benefit from the integration of AI, which will increase their productivity.
In other cases, AI will have the ability to perform core tasks currently performed by humans. This can reduce demand for workers, affect wages and even destroy jobs.
Meanwhile, the IMF predicts that technology will affect just 26% of jobs in low-income countries.
This echoes a report by Goldman Sachs in 2023, which estimated that AI could replace the equivalent of 300 million full-time jobs – but said there could be new jobs along with a boost in productivity.
Meanwhile, UK Prime Minister Rishi Sunak said in November that people shouldn't worry about AI's impact on jobs, as reforms to education will boost skills.
“Many of these countries lack the infrastructure or skilled workforce to harness the benefits of AI, increasing the risk that the technology will worsen inequality between countries over time,” Ms Georgieva said.
More generally, higher-income and younger workers will see an increase in their wage inequality after adopting AI.
Low-income and older workers may hold back, the IMF believes.
“It is critical that countries establish comprehensive social safety nets and provide retraining programs for vulnerable workers,” Ms. Georgieva said. “By doing so, AI can make the transition more inclusive, protect livelihoods and curb inequality.”
The IMF analysis comes as global business and political leaders gather at the World Economic Forum in Davos, Switzerland.
AI is a topic of discussion following the popularity of apps like ChatGPT.
The European Parliament will vote on AI legislative proposals earlier this year, but no legislation will come into effect until at least 2025.
The US, UK and China have yet to issue their own AI guidelines.