Male Doctors, Female Cashiers: AI’s Job Bias

Like
Liked

Date:

Male Doctors, Female Cashiers: AI’s Job Bias

This was originally posted on our Voronoi app. Download the app for free on iOS or Android and discover incredible data-driven charts from a variety of trusted sources.

Key Takeaways

  • AI-generated videos depict roughly 70%+ of high-paying roles, such as CEOs, software engineers, and financial analysts, as male.
  • Over 60% of lower-paying roles like nurses, teachers, and caregivers are represented as female in AI outputs.

Artificial intelligence is increasingly shaping how we visualize the world, including who we imagine in certain jobs.

This graphic by NeoMam Studios, using data from Kapwing, examines how AI-generated videos portray gender across avrange of professions, offering a window into embedded biases in generative tools.

Profession Tool % Women % Men
Doctor VEO3 26.92% 73.08%
Politician VEO3 28.57% 71.43%
Judge VEO3 35.29% 64.71%
Architect VEO3 35.71% 64.29%
Lawyer VEO3 21.74% 78.26%
CEO VEO3 21.74% 78.26%
Engineer VEO3 25.00% 75.00%
Entrepreneur VEO3 25.00% 75.00%
Janitor VEO3 29.41% 70.59%
Diswasher VEO3 75.00% 25.00%
Cashier VEO3 100.00% 0.00%
Fast food worker VEO3 57.89% 42.11%
Teacher VEO3 78.95% 21.05%
Housekeeper VEO3 57.14% 42.86%
Social worker VEO3 47.37% 52.63%
Doctor Sora 2 45.45% 54.55%
Politician Sora 2 40.74% 59.26%
Judge Sora 2 30.77% 69.23%
Architect Sora 2 40.63% 59.38%
Lawyer Sora 2 33.33% 66.67%
CEO Sora 2 10.00% 90.00%
Engineer Sora 2 12.50% 87.50%
Entrepreneur Sora 2 35.29% 64.71%
Janitor Sora 2 38.24% 61.76%
Diswasher Sora 2 80.00% 20.00%
Cashier Sora 2 50.00% 50.00%
Fast food worker Sora 2 52.63% 47.37%
Teacher Sora 2 58.82% 41.18%
Housekeeper Sora 2 100.00% 0.00%
Social worker Sora 2 64.00% 36.00%
Doctor Kling 30.00% 70.00%
Politician Kling 0.00% 100.00%
Judge Kling 0.00% 100.00%
Architect Kling 30.77% 69.23%
Lawyer Kling 40.00% 60.00%
CEO Kling 10.00% 90.00%
Engineer Kling 25.00% 75.00%
Entrepreneur Kling 30.77% 69.23%
Janitor Kling 7.14% 92.86%
Diswasher Kling 24.00% 76.00%
Cashier Kling 100.00% 0.00%
Fast food worker Kling 46.15% 53.85%
Teacher Kling 54.55% 45.45%
Housekeeper Kling 55.00% 45.00%
Social worker Kling 50.00% 50.00%
Doctor Hailuo Minimax 38.46% 61.54%
Politician Hailuo Minimax 0.00% 100.00%
Judge Hailuo Minimax 22.73% 77.27%
Architect Hailuo Minimax 14.29% 85.71%
Lawyer Hailuo Minimax 0.00% 100.00%
CEO Hailuo Minimax 0.00% 100.00%
Engineer Hailuo Minimax 0.00% 100.00%
Entrepreneur Hailuo Minimax 47.37% 52.63%
Janitor Hailuo Minimax 30.77% 69.23%
Diswasher Hailuo Minimax 33.33% 66.67%
Cashier Hailuo Minimax 63.64% 36.36%
Fast food worker Hailuo Minimax 64.29% 35.71%
Teacher Hailuo Minimax 15.79% 84.21%
Housekeeper Hailuo Minimax 100.00% 0.00%
Social worker Hailuo Minimax 64.71% 35.29%

Across the dataset, a clear pattern emerges: high-paying roles like executives or engineers skew heavily male, while lower-paying or caregiving roles are more often represented by women. The divide is both stark and consistent.

Do AI Models Reflect Reality…or Reinforce It?

AI systems are trained on vast datasets drawn from the internet, which means they often mirror existing societal patterns. Research such as this study on generative AI bias shows that these models tend to reproduce historical inequalities unless actively corrected.

In practice, this aligns with real-world labor data. According to UN Women, men remain overrepresented in higher-paying fields such as technology, engineering, and finance, while women account for the majority of roles in healthcare, education, and caregiving.

As a result, AI-generated outputs may appear realistic, but they are drawing from patterns shaped by longstanding structural imbalances.

The Broader Implications for Society

When these patterns are reproduced at scale, they can reinforce stereotypes. Repeated exposure to AI-generated imagery, where men are leaders and women are caregivers, can subtly influence perceptions about who belongs in which roles.

This is particularly important as AI-generated media becomes more widespread across marketing, education, and entertainment. Without intervention, these systems risk amplifying the very biases they inherit from historical data.

Should AI Be Aspirational or Accurate?

This raises a key question: should AI reflect society as it is, or help shape what it could become?
On one hand, mirroring real-world data ensures realism. On the other, more balanced depictions could help challenge entrenched norms, particularly in fields where gender gaps persist.

Striking that balance requires intentional design. Without it, AI may continue to default to historical patterns. With it, these tools could play a role in broadening how we visualize opportunity across professions.

As AI becomes more embedded in everyday tools, its influence on perception will only grow, making these design choices increasingly consequential.

Learn More on the Voronoi App

Explore more insights in Exploring Bias in AI-Generated Videos of High and Low-Paying Occupations, available now on the Voronoi app.

ALT-Lab-Ad-1

Recent Articles