A recent study published in Nature reveals that women are depicted as younger than men across a wide range of online platforms and artificial intelligence systems, despite real-world data showing no significant age differences between genders in the workforce over the past decade. The research analyzed 1.4 million images and videos from sources such as Google, Wikipedia, IMDb, Flickr, and YouTube, along with nine large language models trained on billions of words.
“This kind of age-related gender bias has been seen in other studies of specific industries, and anecdotally, such as in reports of women who are referred to as girls,” said Berkeley Haas Assistant Professor Solène Delecourt, co-author of the study with Douglas Guilbeault of Stanford’s Graduate School of Business and Bhargav Srinivasa Desikan from the University of Oxford/Autonomy Institute. “But no one has previously been able to examine this at such scale.”
The findings show that women are consistently portrayed as younger than men across 3,495 occupational and social categories. This bias was most pronounced in high-status and high-earning jobs. The researchers also found that mainstream algorithms amplify this pattern; when ChatGPT generated nearly 40,000 resumes for various occupations, it assumed women were younger by an average of 1.6 years and had less work experience compared to resumes for men. Additionally, older male applicants were rated as more qualified.
“Online images show the opposite of reality. And even though the internet is wrong, when it tells us this ‘fact’ about the world, and we start believing it to be true,” Guilbeault said. “It brings us deeper into bias and error.”
The team used multiple methods to assess gender and age representation—including human judgment, machine learning tools, and objective information like birthdates—to ensure accuracy across datasets. In all cases, women were linked more strongly with youth while men were associated with older ages.
Delecourt explained: “One concern people might have is that images and videos are kind of unique in that people can wear makeup or apply filters, using image-specific strategies to make themselves look younger. That’s why we also looked at text, and we found exactly the same pattern.”
To test how these biases affect perceptions offline, researchers conducted experiments where participants exposed to occupation-related images online estimated lower average ages for jobs shown with female imagery versus those shown with male imagery. For female-dominated roles, ideal hiring ages recommended by participants skewed younger; for male-dominated roles, they skewed older.
A second experiment involved prompting ChatGPT (gpt-4o-mini) to generate thousands of resumes using matched male and female names. The system routinely depicted women as having more recent graduation dates and less experience.
“These misrepresentations feed directly into the real world in ways that could be widening gaps in the labor market and skewing the ways we associate gender with authority and power,” Delecourt said.
Guilbeault added: “This is of particular concern given the internet is increasingly how we learn about the social world… Our study shows that they are reinforcing stereotypical expectations about how the world should be.”
The study highlights a feedback loop where biased portrayals online reinforce stereotypes among users—potentially affecting hiring decisions—and AI systems trained on these depictions further perpetuate inaccuracies.
“Overall, our study shows that age-related gender bias is a culture-wide, statistical distortion of reality, pervading online media through images, search engines, videos, text…and generative AI,” Delecourt concluded.
The project received funding from The Fisher Center for Business Analytics; The Center for Equity, Gender, and Leadership; The Barbara & Gerson Bakar Fellowship; and The University of California, Berkeley.
