AI could lift productivity 3.8 per cent: IMF
Washington, Sep 26
Artificial intelligence could increase aggregate labour productivity by as much as 3.8 per cent over the long term, according to a new International Monetary Fund working paper that examined patent and employment data across developed economies.
The study found that the rapid expansion of AI-related innovation had already produced measurable productivity gains. It estimated that AI patent activity between 2000 and 2017 increased output per worker by between 0.8 per cent and 1.2 per cent.
The findings were published on Friday in an IMF working paper titled "Artificial Intelligence and Aggregate Labor Productivity: Evidence from Patent Data."
The researchers examined the aggregate effect of AI technology on labour productivity using patent data from countries belonging to the Organisation for Economic Co-operation and Development.
The study covered the period between 2000 and 2017. It found that the issuance of patents connected to AI technology accelerated after 2000 and had more than tripled by 2017.
OECD countries accounted for about 89 per cent of all AI-related patents issued worldwide during the period, according to the paper.
Using a production-function approach, the researchers assessed whether the growing pace of AI patent applications was reflected in increased output per worker.
Their estimates suggest that AI innovation could deliver considerably larger productivity gains in the future as technological development gathers speed and spreads into more industries.
The potential increase of as much as 3.8 per cent is a long-term estimate. The paper indicates that such gains would depend on the continued development and wider adoption of AI technologies across different sectors of the economy.
The benefits are also unlikely to be distributed evenly across countries or employment categories.
According to the study, countries with a high proportion of employees working in professional and managerial positions experienced larger productivity gains from AI technology.
The finding suggests that AI may generate stronger economic returns when it complements the work performed by skilled employees rather than merely replacing existing tasks.
However, those benefits may not appear immediately. Workers need time to learn how to use new systems and incorporate them effectively into their daily responsibilities, the paper said.
The speed with which employees and businesses adapt to AI could therefore determine how rapidly the technology translates into higher output.
Labour-market flexibility could also influence the scale of productivity gains. Greater worker mobility may allow employees to move more easily into occupations and industries where AI can be used most effectively.
The researchers said their conclusions remained robust after accounting for a measure of technological spillovers between countries. Such spillovers occur when AI advances developed in one country are adopted or adapted elsewhere.
The findings underline the connection between innovation, workforce composition and economic policy. Patent creation alone may not be sufficient to produce the largest possible economic gains unless countries also have workers capable of using the technology and labour markets that support occupational mobility.
India has placed increasing policy emphasis on AI, digital public infrastructure and technology-led economic growth. Its large information-technology industry and pool of skilled workers could make productivity, training and workforce mobility particularly important issues as businesses expand their use of AI.
The international debate over AI has increasingly moved beyond technological capability to its effects on jobs, wages and economic growth. Governments are examining how investment in skills, research and digital infrastructure can help capture productivity gains while managing disruption for workers and industries.
— IANS
Leave a comment