According to U.S. Census data, there have been no significant age differences between men and women in the workforce over the past decade, and women, on average, live longer than men. Yet, this balance isn’t reflected in what we see across the internet—or even in the way artificial intelligence systems describe people.
A new study published in Nature uncovers a systematic digital bias: women are consistently depicted as younger than men. The research team—led by Solène Delecourt (Berkeley Haas), Douglas Guilbeault (Stanford Graduate School of Business), and Bhargav Srinivasa Desikan (University of Oxford/Autonomy Institute)—analyzed 1.4 million online images and videos as well as data from nine large language models (LLMs) trained on billions of words. Their analysis spanned major platforms such as Google, Wikipedia, IMDb, Flickr, and YouTube, and covered 3,495 occupational and social categories. Across nearly all of them, women were portrayed as younger than their male counterparts.
“This kind of age-related gender bias has long been noted in specific industries and cultural narratives for example, in the way adult women are often referred to as ‘girls,’” says Delecourt. “But this is the first time anyone has been able to document such a bias at a global, data-driven scale.”
The bias was most pronounced in high-status, high-income professions, where women were especially likely to appear younger than men in similar roles. Even more concerning, the researchers found that mainstream AI systems amplify these stereotypes. In tests involving nearly 40,000 AI-generated résumés, models such as ChatGPT assumed women were younger and less experienced, while rating older male candidates as more qualified.
As co-author Guilbeault explains, “Online content paints a version of reality that’s the opposite of what’s true. And even though the internet is wrong, when it repeatedly shows us these images and narratives, we start to believe them. That’s how digital bias quietly shapes our perceptions and, ultimately, deepens inequality.”
This large-scale study highlights how pervasive and subtle gendered ageism can be in the digital world, influencing not just how people are portrayed online, but also how AI systems learn, replicate, and reinforce these biases.
Source: https://newsroom.haas.berkeley.edu/news-release/women-portrayed-as-younger-than-men-online-and-ai-amplifies-the-bias/
