Artificial Intelligence Growth At the 2023 World Economic Forum, technology entrepreneur Mihir Shukla remarked: “People keep saying AI is coming, but it is already here.” Over the past decade, the use of artificial intelligence (AI) in everyday activities has surged dramatically, and ChatGPT (developed by OpenAI) stands as a prime illustration. This widely adopted generative AI tool is now used by more than a billion people for daily tasks such as programming and content creation. The pace and scale of adoption are striking: ChatGPT reached 100 million users within just 60 days, whereas Instagram needed two years to achieve the same number. According to Artificial Intelligence Growth a recent report from Stanford University, the number of AI-related patents grew thirtyfold between 2015 and 2021 (HAI 2023), underscoring the extraordinary speed of AI innovation. Today, AI-powered systems can handle a wide array of functions—information retrieval, logistics coordination, financial transactions, document translation, business writing, legal brief preparation, and even medical diagnoses. These tools are expected to enhance both precision and efficiency since they continually improve through machine learning (ML).

AI is widely seen as a driver of economic growth and productivity. By analysing and processing massive amounts of data, AI has the capacity to optimize business processes and raise efficiency. The McKinsey Global Institute forecasts that by 2030, Artificial Intelligence Growth roughly 70% of firms will implement at least one form of AI, though fewer than half of large enterprises are expected to utilize AI across all domains. Meanwhile, PricewaterhouseCoopers estimates that AI could expand global GDP by 14% by 2030 (PwC 2017).

Read more about Budget friendly car

Read more about Artificial intelligence growth

Scholarly work on AI’s implications for the job market has grown in recent years. Acemoglu and Restrepo (2018) developed a framework to examine how emerging technologies affect employment. They divide the impact into three major channels: the displacement effect (when machines replace workers), the productivity effect (when technology boosts efficiency and output), Artificial Intelligence Growth and the reinstatement effect (when innovations enable new kinds of tasks, particularly in services where human skills still hold an advantage, thereby increasing labour demand).

Frank et al. (2019) divide the existing literature on AI’s impact on the labour market into two main strands: the pessimistic viewpoint and the optimistic viewpoint. Pessimists argue that AI-driven labour substitution will damage employment prospects. Frey and Osborne (2013) projected that nearly 47% of jobs in the United States could be automated within a decade. Their study showed that a large share of roles in service industries – which have been the main driver of U.S. job growth in recent decades Artificial Intelligence Growth – are particularly vulnerable to computerisation. Building on this framework, Bowles (2014) estimated that 54% of employment in the European Union faces similar risks. Likewise, Acemoglu and Restrepo (2017) present historical evidence of adverse labour market effects from automation, showing that U.S. regions most exposed to industrial automation during the 1990s and 2000s experienced significant, persistent declines in both employment and wages, due to weak productivity and reinstatement effects.

AI is also expected to reshape the structure of employment. Autor (2015) documented a trend of labour market polarisation over the past few decades, with growth in high- and low-skilled jobs but declines in middle-skilled ones as computers replaced routine work. However, he suggested that this trend could eventually reverse: certain low- and middle-skilled jobs may be relatively resistant to automation, while some highly skilled but routine-based roles might be more susceptible to new technologies such as AI. In contrast, Petropoulos and Brekelmans (2020) argued that the AI revolution, unlike the computerisation and robotics wave, is unlikely to drive job polarisation, since it has the potential to affect occupations across low-, medium-, and high-skilled categories.

Supporters of the optimistic view contend that AI’s productivity and reinstatement effects will outweigh its substitution effects. Some forecasts suggest that AI and robotics could create as many as 90 million jobs by 2025, implying a net positive labour market effect. In line with this, the World Economic Forum projected in October 2020 that although 85 million jobs may disappear globally due to AI by 2025, around 97 million new roles would emerge in areas such as data science, machine learning, cybersecurity, and digital marketing.

Lawrence et al. (2017) maintain that AI automation is unlikely to harm the labour market because of strong positive spillover effects (the reinstatement effect), which offset direct substitution losses. They describe this process as a form of Schumpeterian ‘creative destruction’, with automation reshaping rather than eliminating work. In a similar vein, Arntz et al. (2016) estimated that only 9% of jobs in the UK face automation risks within the next decade. They argue that job transformation is more likely than job loss, with 35% of positions expected to undergo significant changes over the next twenty years.

Nakamura and Zeira (2018), using a task-based theoretical framework, show that automation does not necessarily result in long-term unemployment. A systematic review by Somers et al. (2022) of empirical studies on technological change and employment finds that although many studies highlight labour substitution effects, they are outweighed by those documenting job creation, reinstatement, and real-income effects. Their conclusion is that the overall employment impact of technology tends to be positive rather than negative. Supporting this, Bholat (2020) points out that while specific industries often experience job losses from new technologies, these effects have historically been offset by broader gains in real income. Cheaper, higher-quality goods and services boost disposable income, which fuels demand for new products and subsequently generates employment in other sectors. Alan Manning also notes (cited in Bholat, 2020) that some of the most pessimistic forecasts about automation’s impact on jobs over the past decade have not materialised, suggesting that fears about AI-driven unemployment may be somewhat overstated.

In May 2023, the CfM-CEPR survey asked its panel members to predict the influence of AI on global economic growth and unemployment in high-income economies over the next ten years. The first question concerned AI’s projected effect on global growth, while the second focused on its expected impact on unemployment rates in advanced economies.

Twenty-seven experts took part in this survey. Most of them (64%) expect AI to push global economic growth to around 4–6% annually over the next ten years. The rest (36%) anticipate that AI will not have a noticeable effect on overall growth. Importantly, many respondents admitted to being highly uncertain about their forecasts.

Nearly two-thirds of the participants anticipate that AI’s progress in the coming decade will have a positive impact on economic performance. Jorge Miguel Bravo (Nova School of Business and Economics, Lisbon) points to the spread of machine learning as a “general purpose technology” across both advanced and emerging economies, saying this makes him “bet on the upside rather than no change or decline.” Ugo Panizza (The Graduate Institute, Geneva (HEID)) shares a similar perspective, asserting that “AI will raise productivity and therefore foster stronger economic growth.” Still, he highlights potential downsides, warning that “AI could raise unemployment and inequality, which might feed back negatively into productivity and growth.” Robert Kollmann (Université Libre de Bruxelles) also voices cautious optimism, contending that while AI is unlikely to alter the long-run global GDP growth trend, it might deliver a modest boost of about “0.5 percentage points.”

A large portion of the group stresses that the effect of AI on global growth is deeply uncertain, making reliable predictions unrealistic. Andrea Ferrero (University of Oxford) reflects this sentiment: “I expect AI to matter for the economy, but I cannot clearly foresee how. Some industries may thrive, while others might struggle. Overall consequences for growth remain very uncertain in my view.” Jagjit Chadha (National Institute of Economic and Social Research) remarks that the eventual impact of AI hinges on multiple elements – “which policy choices are made, how monopoly power is checked, and what innovations emerge” – concluding that he cannot forecast the outcome “with any degree of certainty.”

Ricardo Reis (London School of Economics) concisely captures the panel’s shared hesitation: “Making decade-ahead growth projections is extremely difficult, so ‘not confident’ is the most important takeaway from this answer.”

Twenty-nine experts took part in this survey question. A majority (63%) believe that artificial intelligence will have little to no impact on employment levels in advanced economies over the next ten years. Of the remaining panellists, most (27%) expect that AI developments could push unemployment higher in these countries. Only two members of the panel anticipate that AI might instead reduce unemployment during this period. Importantly, over half of the respondents admitted low confidence in their answers, underscoring the significant uncertainty surrounding this issue.

Looking further ahead, most panellists expect AI to leave joblessness in high-income economies largely unaffected. Michael Wickens (Cardiff Business School and University of York) points to historical examples of technological change to support his position: “Employment and unemployment will remain stable, but working hours will decline and leisure time will expand. This has been the outcome of past technological progress, and AI will follow the same path.” Similarly, Cédric Tille (The Graduate Institute, Geneva) contends that “unemployment effects may be modest,” though he stresses that “the consequences for inequality and the need for redistribution could be considerable.” Maria Demertzis (Bruegel) argues that the effect on unemployment will hinge on reskilling, explaining: “The faster reskilling takes place, the smaller the employment impact will be.”

Some respondents emphasize the unpredictability of AI’s long-run impact on work. Andrea Ferrero (University of Oxford) illustrates this perspective: “I anticipate a short-term rise in unemployment as certain roles are replaced by machines. Over time, labor supply will adjust, and AI could even help reduce the natural unemployment rate—but I cannot say for certain.”

A smaller share of the panel voices concern that AI will harm groups already disadvantaged in the labor market. Wouter den Haan (London School of Economics) captures this worry: “I fear that AI will negatively affect vulnerable workers, especially those who struggle to adapt to new circumstances.”

By contrast, a handful of panellists suggest that AI could actually reduce unemployment in developed economies. Volker Wieland (Goethe University Frankfurt and IMFS) proposes that AI may lower both unemployment rates and the number of hours worked.