To benefit from AI, companies must train all employees, support certain interest groups, and possibly re-organize work around AI.
The introduction of generative AI to the market is undoubtedly a significant event in the history of information technology. Around the world’s technology hubs, it has been taken for granted that AI will fundamentally change the picture of the economy, or it has already done just that. But for AI to fully live up to its expectations, companies must first acquire this technology and then become more productive by really using it.
The combined market value of the big five tech companies (Alphabet, Amazon, Apple, Meta, and Microsoft) increased by about two trillion dollars over the past year. Translated into expectations, this would mean a roughly 300 to 400 billion US dollars increase in annual revenues and roughly a share equivalent to a sixth tech giant’s revenue. For the time being, the five tech giants’ performance lags far from expectations. Even the most confident analysts estimate that a mere 10 billion of Microsoft’s earnings will come from generative AI-related business this year.
The stock market index, which tracks the companies estimated to benefit the most from AI and therefore have the most AI-driven growth potential through increased productivity, shows that since the end of 2022, the stock market returns of such companies (e.g., Walmart and H&R Block) have not performed any better than average, but quite the opposite.
Explanations can be sought, for example, in challenges related to generative AI in working life, companies’ reluctance to invest in a rapidly developing and, therefore, permanently unfinished technology, or in the fragmentation of AI investment into many small experiments. However, a common factor seems to be the slow uptake of AI, leading to weaker-than-anticipated productivity development.
According to Microsoft’s own estimates, up to 75% of global information workers use artificial intelligence in their work. However, a survey commissioned by the U.S. Census Bureau, recognized for its independent and reliable analyses, shows that only 5% of American companies have utilized AI two weeks before their response to the survey.
Economists Anders Humlum and Emilie Vestergaard from Chicago and Copenhagen conducted an extensive study of approximately 100,000 Danish respondents to investigate the use of ChatGPT and its impact on work productivity.
The study included 11 occupational groups, accounting for approximately 21% of the Danish workforce. The occupational groups studied were accountants, customer support specialists, financial specialists, HR professionals, IT support specialists, journalists, lawyers, marketing specialists, office clerks, software developers, and teachers. Between these occupational groups, AI adoption rates ranged between 34% for financial experts and 79% for software developers.
Until now, the decision to use artificial intelligence at work has typically been a personal choice for each employee. Companies could do more to facilitate and accelerate the use of AI by providing support and training on how to utilize AI. Employees with less experience and expertise may need additional assistance to benefit from generative AI.
Humlum and Vestergaard’s study also revealed significant differences in genders’ AI use. Women are about one-fifth less likely to use ChatGPT than men in the same occupational group. A targeted training program could help to reduce the significant gender gap.
Many survey participants indicated that they do not seek to increase the performance of tasks where ChatGPT can save time by changing tasks.
This phenomenon was particularly evident among teachers, suggesting that to increase overall productivity, companies should redesign and reorganize work in accordance with generative AI, taking into account its opportunities and limitations. Only then can time saved with the help of artificial intelligence help increase production and improve profitability.
Sources:
What happened to the artificial-intelligence revolution? (economist.com)
The Adoption of ChatGPT by Anders Humlum, Emilie Vestergaard :: SSRN
Henry Scheinin
Leading consultant, Partner
