ChatGPT is widespread in the exposed occupations, but inequalities have emerged. Women are 20 percentage points less likely to use the tool.
The arrival of ChatGPT marks the era of Generative Artificial Intelligence (AI), where intelligent algorithms have the potential to reshape various high-skilled professions. Our IZA Discussion Paper provides both descriptive and experimental insights into the adoption of ChatGPT. We explore who has embraced this technology, how workers perceive its impact on their jobs, and the factors influencing its adoption.
Teaming up with Statistics Denmark, we conducted a survey involving 100,000 workers across eleven different occupations. These include accountants, customer support specialists, financial advisors, HR professionals, IT support specialists, journalists, legal professionals, marketing professionals, office clerks, software developers, and teachers. Our survey includes an experiment, informing workers about expert assessments of ChatGPT in their job tasks.
Who is using ChatGPT? ChatGPT has gained considerable attraction among workers in these professions, with half of them having utilized the technology. Adoption rates vary, with software developers leading at 79% and financial advisors trailing at 34%. Notably, there is a significant gender gap, with women lagging behind men by 20 percentage points in usage. Interestingly, despite being younger and less experienced, early adopters of ChatGPT were already earning slightly more before its introduction. These adoption trends hint at the necessity of providing additional support to women and less accomplished workers to fully leverage the benefits of Generative AI.
What do workers believe about ChatGPT? Professionals in these fields recognize the substantial productivity gains afforded by ChatGPT, estimating that it could halve the time required for a third of their tasks. They acknowledge that these time savings stem from ChatGPT substituting for human expertise. However, beliefs regarding the technology's efficacy vary significantly among workers within the same occupation. This variability underscores the inherent uncertainty surrounding general-purpose technologies like ChatGPT, which boast broad and unpredictable applications.
Do belief disparities influence adoption? We explore whether workers' beliefs influence their adoption of ChatGPT by exposing a random subset of workers to expert assessments of its time-saving potential. This informational intervention effectively alters workers' perceptions, a shift that persists even in a follow-up survey conducted two weeks later. However, despite these changed beliefs, there is limited impact on actual adoption rates.
What prevents workers from using ChatGPT? Workers cite employer restrictions and the need for training as the primary hurdles to embracing ChatGPT. Interestingly, few express concerns about existential threats, such as over-reliance on technology or job obsolescence.
Looking ahead, our findings underscore the importance of corporate policies, such as providing usage guidelines and facilitating training, in encouraging wider ChatGPT adoption. By doing so, these measures can also help address the gender and income disparities evident in its usage.
© Anders Humlum and Emilie Vestergaard
Anders Humlum is Assistant Professor of Economics at the University of Chicago, Booth School of Business.
Emilie Vestergaard is PhD Student at the University of Copenhagen
Please note:
We recognize that IZA World of Labor articles may prompt discussion and possibly controversy. Opinion pieces, such as the one above, capture ideas and debates concisely, and anchor them with real-world examples. Opinions stated here do not necessarily reflect those of the IZA.
Related IZA World of Labor content:
How is new technology changing job design? By Michael Gibbs Sergei Bazylik
Who owns the robots rules the world by Richard B. Freeman
Interaction between technology and recruiting practices by Vera Brencic
Photo by Growtika on Unsplash