Up Next

Market-Based Solutions to Vital Economic Issues


Kenan Institute 2024 Grand Challenge: Business Resilience
Market-Based Solutions to Vital Economic Issues
Kenan Insight
Apr 18, 2023

Will Generative AI Disproportionately Affect the Jobs of Women?

Technology is neither good nor bad; nor is it neutral.

Kranzberg’s First Law of Technology

The release of generative artificial intelligence products such as ChatGPT, Bing Chat, Bard, Midjourney and others – which allow anyone to use prompts to generate text, images, music or video for business or personal use – have democratized AI. By doing so, generative AI has become a game changer for many industries, offering new ways to automate tasks, increase productivity and improve quality.

While this technology will create new job opportunities and a growth in GDP, the exposure of generative AI to automate tasks in existing jobs will have an impact on those occupations as well. These include changes in job tasks and professional roles, the need to learn new skills to remain competitive and, unfortunately, the loss of jobs.

It may also mean, however, that workers will be freed up to be more efficient and creative, thus increasing both productivity and quality. New jobs will be created as well, and workers may find more meaning in their work if AI allows tasks that are time-consuming to be automated. Therefore, “impact” should not be thought of strictly as a negative, since there will also be positive changes in these job categories.

This analysis aims to analyze whether the potential impact of generative AI on jobs will impact men and women differently.

Main Finding

The main finding of this analysis is that eight out of 10 women (58.87 million) in the U.S. workforce are in occupations highly exposed to generative AI automation (more than 25% of the occupational tasks) vs. six out of 10 men (48.62 million). Overall, 21% more women are exposed to AI automation than men even though men outnumber women in the workforce. This is due to the affected occupations being populated by more women than men. “Highly exposed” means 25%-50% of the tasks in that occupation could be automated by generative AI.

This finding was determined using Goldman Sachs’ “The Potentially Large Effects of Artificial Intelligence on Economic Growth” as a base. The report identified 15 occupations that would be most affected. Within those occupations, the total number of employees and the gender breakdown determined the total number of men and women exposed to generative AI automation.

Frontiers of Business Conference: Workforce Disrupted

Our 2023 Frontiers of Business Conference will convene corporate executives, top researchers and policy leaders to share objective, evidence-based solutions for navigating the precarious road toward a labor market equilibrium. Learn more today.

Analysis and Detailed Findings

The Goldman Sachs report had the following key takeaways:

  • About two-thirds of jobs will be impacted by generative AI. Based on the employment share of those occupations and that 25%-50% percent of tasks in those jobs are exposed to automation, it is projected that up to a quarter of the work tasks being performed today could be automated by generative AI.
  • Taking estimates from U.S. and European analysis worldwide, generative AI may potentially replace the equivalent of 300 million full-time jobs.
  • However, the automation of those tasks is not the only potential outcome. If history is a guide, these job losses will be counterbalanced by the formation of new positions, and new types of jobs are likely to appear.
  • These changes may also result in significant increases in annual global GDP, up to 7%, with this increase in economic activity generating demand for new goods and services that is likely to spur job creation.

More specifically, the Goldman Sachs report projected that the following 15 occupations would be the most affected, with 25%-50% of the tasks in those occupations potentially exposed to automation via generative AI (see Figure 1). As you can see, these are primarily knowledge worker positions. Generative AI will have much less impact on jobs that require physical labor as, at this point in time, it lacks ability to perform that type of work.

Figure 1: Occupations and Task Exposure to Automation via Generative AIMethodology: Gender Breakdown Analysis

Office and Administrative Support46%
Architecture and Engineering37%
Life, Physical, and Social Science36%
Business and Financial Operations35%
Community and Social Service33%
Sales and Related31%
Computer and Mathematical29%
Farming, Fishing, and Forestry28%
Protective Service28%
Healthcare Practitioners and Technical28%
Educational Instruction and Library27%
Healthcare Support26%
Arts, Design, Entertainment, Sports, and Media26%

Using the Goldman Sachs list of the top 15 job categories impacted by generative AI (above), we then analyzed the gender breakdown for these occupations based on the U.S. Occupational Employment and Wage Statistics  Report. These data come from the Bureau of Labor and Statistics, the source used in the Goldman Sachs report to determine the employment share of each occupation. The gender breakdown of the total of all the top 15 occupations were then calculated to determine the impact of generative AI by gender.

Detailed Findings

Figure 1 below shows the female-to-male percentages for the 15 occupations, ranked in order of number of employees in each occupation, largest to smallest. By looking at the total at the bottom of the chart, one can see there are significantly more women in these occupations (58.87M) than men (48.62M) even though there are more men (84.21M) in the workforce than women (74.08M). Thus almost 80% of women in the workforce are in occupations exposed to automation via generative AI vs. 58% of men.

Figure 2: Detailed View of Gender Breakdown of 15 Occupations Most Exposed to Automation from Generative AI

Gender breakdown of jobs most heavily impacted by AI

Management occupations20.20M40.5%59.5%
Office and administrative support occupations16.10M71.9%28.1%
Sales and related occupations14.32M49.4%50.6%
Healthcare practitioners and technical occupations9.81M75.7%24.3%
Education, training, and library occupations9.22M73.3%26.7%
Business and financial operations occupations9.15M54.5%45.5%
Computer and mathematical occupations6.17M26.7%73.3%
Healthcare support occupations4.93M84.6%15.4%
Architecture and engineering occupations3.46M16.1%83.9%
Arts, design, entertainment, sports, and media occupations3.44M48.8%51.2%
Protective service occupations3.06M23.2%76.8%
Community and social services2.95M67.2%32.8%
Legal occupations1.86M52.6%47.4%
Life, physical, and social science occupations1.84M48.2%51.8%
Farming, fishing, and forestry occupations.98M26.2%73.8%
Total of all jobs158.29M74.08M84.21M
Total of jobs impacted / total jobs68%79%58%

The reason more women than men are exposed to AI automation is straightforward: A higher percentage of working women are in white-collar jobs (~70%) vs. blue-collar ones (~30%) while for men the ratio is roughly 50/50.


As discussed earlier, “impacted by generative AI” may mean positive or negative changes for society and workers. For society, it may bring economic growth as well as societal disruption. Whether the changes are good or ill for individual workers will depend on their occupation, firm, individual capabilities and ability to adapt. Some will adjust better than others. There will be winners and losers.

For every knowledge worker, there are some skills and attitudes that will enable them to adapt and thrive in the coming AI era. The first is a need to gain generative AI capabilities. As the saying goes: “You won’t be replaced by AI. You will be replaced by someone who knows AI.” Cognitive workers need to become familiar and then fluent in using generative AI tools and applying them in their current jobs to become more productive, knowledgeable and creative. They should also remain well informed about AI trends and how these tools are being used in their professions, industries and firms so they can find additional means of improving their value to their organizations.

Also, as New York Times columnist David Brooks put it, “In the age of AI, major in being human.” There are things humans can do that AI cannot do well or at all. These include taking the initiative, discovering problems, influencing others, developing creative solutions and navigating organizational politics – all things that are crucial to getting things done in organizations.

Lastly, resiliency and adaptability will be crucial. AI will move ever faster and disrupt both the workplace and society. Being able to cope with and respond positively to those changes will set one apart from the rest.

What should society do if the impact of AI in the workplace falls more on women than men is a question we must grapple with as we see the changes it brings. As discussed, the results will be a mix of positives and negatives, and at this point, it is difficult to say what the balance of those will be. While that may not be a satisfactory answer, we will have to adjust as the effects of generative AI develop.

Considerations and Limitations

  • The record adoption of ChatGPT, followed by a proliferation of other generative AI, has been both very recent and striking. It is still early in our understanding of how these products will change our future. This fact, compounded by the release of a wider array of increasingly powerful products on a seemingly weekly basis, means any analysis of their future impact must be taken as only an early projection of what may happen.
  • Secondly, the occupation categories are quite broad. Within those job categories, some specific subcategories will be more affected than others. Those subcategories may have different male-to-female ratios than others. For example, in the legal profession, 52.6% of employees are women. However, the paralegal subcategory, which will be more likely to be impacted than other legal subcategories, is composed of 83.2% women. Therefore, future research at the subcategory level would be useful in determining impact at a more detailed level.

Data for this report were compiled by Paige Smith, UNC Kenan-Flagler Business School MBA candidate.

You may also be interested in: