The emergence of generative artificial intelligence (AI) inevitably raises the question of a rapid acceleration of task automation that will increase productivity. Despite the uncertainty that remains about its potential, its ability to generate content that is indistinguishable from that created by man and to remove communication barriers is a major advance that is likely to have considerable macroeconomic effects.
In short, If generative AI keeps its promises, the job market could be strongly reshaped. According to several studies that aggregate data from the United States and Europe, around two-thirds of current jobs would be exposed to some degree of automation by AI. Even more, generative AI could replace up to a quarter of the current job.
However, this perspective of the destruction of low-skilled jobs must be nuanced in the light of historical dynamics. Indeed, job losses due to automation have always been offset by the creation of new jobs. Moreover, the emergence of new professions as a result of technological innovations is responsible for most of the long-term employment growth. The combination of a significant reduction in labor costs, the creation of new jobs, and higher productivity then raises the possibility of a productivity boom that would significantly increase economic growth, although it is difficult to predict when this boom will occur.
Generative AI, a boom explained
The exponential increase in available computing power in recent years has allowed for rapid advances in the complexity of the tasks that AI can perform and in the precision with which it can perform them. For example, the latest iteration of OpenAI's GPT model - GPT-4, released in March 2023, about a year after the completion of the training of the GPT-3.5 model that currently underlies ChatGPT - is 40% more likely to produce accurate responses and can now accept visual data (not just text). The algorithms behind generative AI had in fact begun to go beyond human references for tasks such as image classification and reading comprehension, even before these recent advances.
As AI became more and more advanced and accessible, interest and investments followed. In 2021, U.S. and global private investments in AI were $53 billion and $94 billion respectively - each having increased more than fivefold in real terms compared to the previous five years. While investments continue to increase at the more modest rate of growth in software investments during the 1990s, theUS investments in AI alone could approach 1% of US GDP by 2030.
While there are still many uncertainties about the capacity and timeframe for the adoption of generative AI, these developments suggest that AI is well-positioned to advance rapidly and gain momentum in the years to come.
The future of the labour market: replace sometimes, supplement often
The most recent research suggests that a large portion of employment and work is at least partially exposed to automation by AI, which suggests significant labor savings. The final portion of work exposed to automation could be between 15% and 35%, a range that is often mentioned but could be conservative.
While the impact of AI on the job market is likely to be significant, most jobs and industries are only partially exposed to automation and are therefore more likely to be supplemented than replaced by AI. Sectors where the replacement rate could be greater than 50% are the legal and administrative sectors, which include numerous office jobs with repetitive tasks that are easily replaceable.
What is the impact on productivity and growth?
The significant portion of employment exposed to automation through generative AI increases the potential for a boom in labor productivity. AI-driven automation could increase global GDP in two main ways.
First, most workers are in jobs that are partially exposed to AI automation and, after the adoption of AI, will likely apply at least some of their freed up capacity to productive activities that increase production. University studies confirm that workers at companies that adopted AI at an early stage experience higher growth in labor productivity as a result of the adoption of AI, with estimates generally indicating a increase of 2 to 3 percentage points per year. While it is difficult to extrapolate these results due to differences in generative AI's capacity compared to previous generations, they clearly suggest that generative AI can lead to an economically significant increase in productivity.
Second, it is likely that many workers displaced by AI automation will end up being reused - and therefore stimulating total production - in new professions that will emerge either directly from the adoption of AI, or in response to the higher level of global demand and labor demand generated by the increase in productivity of workers who have not been displaced.
These two paths have numerous historical precedents. For example, innovations in the field of information technology have given rise to new professions such as web designers, software developers, and digital marketing professionals, but they also increased overall income and indirectly boosted demand for service sector workers in industries such as health care, education, and food services.
Technological change has displaced workers and created new job opportunities at about the same rate during the first half of the post-war period, but it displaced workers more quickly than it created new opportunities since the 1980s. These results suggest that the direct effects of generative AI on labor demand could be negative in the short term but lead to a real revolution in the job market in future years.
About StoryShaper:
StoryShaper is an innovative start-up that supports its customers in defining their digital strategy and the development of automation solutions tailor-made.
Sources: StoryShaper, Goldman Sachs.
