AI to replace 40% of productivity? Barclays conducted a detailed calculation

After a year of hype, almost everyone who has come into contact with OpenAI has felt its impact on both life and work.

But from a global perspective, how much value does generative AI really bring to productivity?

According to Kristalina Georgieva, Managing Director of the International Monetary Fund (IMF), artificial intelligence will affect about 40% of global employment, with approximately 60% of jobs in developed economies potentially being impacted by AI. This impact can be both positive and negative: about half of the jobs may benefit from AI technology, leading to increased productivity, while the other half may be replaced by AI, resulting in reduced demand for labor and lower wages. In the most extreme cases, some jobs may disappear.

Is the future really that bleak? Barclays Bank believes that in the context of slow global labor productivity growth, the latest breakthroughs in AI can have a significant impact on employment and industries, benefiting both developed and developing countries.

The Two Core Attributes of AI: Accessibility and Versatility

Artificial intelligence is not a new phenomenon. Its origins can be traced back to Alan Turing's 1936 paper "On Computable Numbers, with an Application to the Entscheidungsproblem." From the 1980s to the present day, breakthroughs in AI technology have always caused a stir in popular culture.

But Barclays believes that the current situation is different. Compared to previous technological advancements, generative AI represented by OpenAI is revolutionary.

Analysts point out that the two core attributes of AI make it capable of bringing long-awaited productivity improvements to the world:

1. Accessibility: AI technology can be used by a wide range of users on existing infrastructure. Any user can issue basic commands to tools like OpenAI without the need to learn any special programming languages. Software providers can bundle AI capabilities into existing search engines and office suites, or even integrate them into user hardware, thereby introducing AI functionality.

2. Versatility: AI tools are not limited to any specific task, function, problem, or department. This makes them applicable to various disciplines. For example, large language models can summarize legal documents, medical files, or insurance documents after being trained on text. These tools are not limited to text either: GenAI systems can combine text and images with videos, audio, and even robotic functions.

Taking these two points into consideration, the bank believes that compared to traditional AI, the return on investment for GenAI is higher.

The return on investment generated by traditional AI is quite limited. According to estimates from the IBM Institute for Business Value, the average return on investment increased from about 1% in early 2020 to about 6% by the end of 2021. Compared to that, the estimated return rate of generative AI in 2022 has already reached 6%, and is expected to climb to 10% or even higher by 2025.

Barclays believes that the successful application of artificial intelligence is likely to follow the "J-curve" pattern typical of transformative technologies, starting slowly but experiencing exponential growth after surpassing a certain threshold.

The greatest value of AI lies in improving human labor productivity.

Barclays believes that as artificial intelligence is more widely applied, it can address some of the most pressing issues facing the global economy, such as population aging in developed countries and the low per capita productivity prospects in developing countries.

Firstly, Barclays points out that the main challenge currently faced by developed countries is the decline in labor productivity caused by an aging workforce, while the main problem faced by developing countries is the limited average skills and education levels of their workforce. AI can effectively solve both of these problems:

For developed countries, economic growth is driven by the quantity of labor used and/or productivity. If an economy can obtain more from its labor reserves and increase output per hour worked, it can offset the loss of capacity caused by aging.

For developing countries, the working-age population is still growing overall, and in some cases, the growth rate is very fast. However, compared with developed economies, emerging economies often have limited average skills and education levels, resulting in lower GDP per hour worked. We acknowledge that the widespread application of AI may put developing economies at a disadvantage as it brings the ability to automate simple services such as call centers. Nevertheless, there is evidence that AI has the greatest positive impact on individuals with basic skills, enabling them to go further. If the wider adoption of AI does indeed lead to an increase in jobs, it can create more productive positions and higher wages.

Barclays believes that to ensure the full potential of AI technology is realized and its negative impacts are limited, the correct policy combination needs to be formulated from a regulatory perspective and at the enterprise level.

Specifically, AI technology should be seen as a complement to the workforce rather than a replacement. The good news is that currently, most executives do not have the idea of completely replacing human labor with AI, but rather view AI as a supplement to human resources:

A recent survey by IBM's Institute for Business Value shows that companies are placing more emphasis on human unique talents, such as creativity, social and interpersonal skills, and empathy. In other words, the increasing use of AI is likely to change the way companies utilize "hard" skills internally and may favor employees who can provide a combination of hard and soft skills.