A new randomized controlled study by Harvard and Stanford economists finds that ChatGPT can significantly increase the productivity of professionals who do everyday writing tasks.

Participants were randomly assigned realistic work simulations with or without access to ChatGPT.

Those who had access to the AI ​​tool completed tasks 40% faster and produced 18% better quality, according to experts.

The study adds a new dimension to debates about how AI will impact jobs and concludes that ChatGPT complements human professionals rather than replaces them.

ChatGPT Productivity Study

The study, by economists Shoji Noy of Harvard University and Zixuan Zhang of Stanford University, assigned more than 450 college graduates realistic writing tasks that resembled their actual work.

Participants included marketers, consultants, data analysts, and other professionals who produce written documents on a regular basis.

The study shows:

“Workers exposed to ChatGPT during the experiment were twice as likely to use ChatGPT at their actual workplace two weeks after the experiment and 1.6 times more likely to be using ChatGPT two months later.”

ChatGPT narrows the skills gap

The study found that ChatGPT reduced inequality between higher-skilled and lower-skilled professionals.

Participants who performed worse on their first writing task improved more when given access to ChatGPT on their second task than the control group.

The study states:

“In the treatment group, initial inequalities were more than half eliminated by treatment. This reduction in inequality was driven by the fact that participants who performed worse on the first task benefited more from ChatGPT access.”

ChatGPT helps people rather than replaces them

The authors say their findings suggest that AI tools like ChatGPT complement rather than replace human professionals.

However, they point out that their brief experiment examined only a limited range of occupations and incentive structures.

“Compared to our experimental results, ChatGPT’s direct productivity effects in the real economy will be somewhat smaller and the technology will be more complementary to human labor.”

The researchers call for further studies on how productivity gains from AI affect wages, employment and changes in demand across the economy.

Their experiment shows that advanced AI tools will have profound effects on companies and labor markets that are difficult to predict.

Further details on the study methodology

The study was a randomized controlled trial that is considered the gold standard for evaluating interventions. Participants were randomly assigned whether or not they had access to ChatGPT.

The writing tasks were designed based on actual activities that professionals do in the workplace, such as writing reports, press releases, emails, etc. Participants rated the tasks as realistic.

The output quality was evaluated by experienced professionals who had no idea about the experimental conditions. Participants were offered incentives for quality work.

Participants received incentives with bonus payments based on the quality of their writing, assessed by experts who were unaware of the study conditions. On average, participants earned $17 an hour.

Diploma

This randomized study shows that ChatGPT can increase productivity. However, the study examined only a limited range of occupations and incentives.

As advanced AI spreads across the economy, it’s important to continue studying its impact on wages, employment, inequality, and more.

While the technology can bring benefits to professionals, the broader economic implications remain complex and difficult to predict.

Featured image: Vitor Miranda/Shutterstock