AI models misestimate mental health stigma across gender groups
AI Representations of Mental Health Stigma: Quantifying the Gap Between AI Estimations and Human Perceptions Across Gender Groups Using GPT-Based LLMs
Computers and SocietyArtificial Intelligence
Summary
AI systems used in mental health sometimes show bias about how different genders feel about mental illness. The authors studied two advanced AI language models, GPT-4 and GPT-5, to see if they represent those feelings accurately. They found the AI models roughly match how men, women, and nonbinary people differ in stigma, but the AI often gets the size of these feelings wrong. This means AI might not fully understand mental health stigma as people experience it, so careful checking is needed before using AI in this area.
What this means in practice
- •For mental health app developers: Adjust AI-driven assessments of mental health stigma to better reflect real human perceptions across genders.
- •For bias audit teams: Use AI-human comparison methods to detect and correct gender-based distortions in AI models applied to mental health.
Authors
Soojong Kim, Hye Min Kim, Kwanho Kim, Dongdong Zhu
Abstract
The increasing use of artificial intelligence (AI) in mental health applications raises concerns about its potential to internalize and perpetuate societal biases. This research evaluates both disparities and misalignment in how large language models (LLMs), GPT-4 and GPT-5, represent perceptions of mental health across gender groups. The research comprises two studies. Study 1 applied widely adopted measures of Social Distance, Stereotypes, and Attitudes Toward Seeking Help to assess whether the AI models depict systematic gender differences. Study 2 compared AI estimations of Personal, Self, and Perceived Stigma with large-scale real-world human data from the Healthy Minds Study (N = 76,406) to evaluate alignment between AI estimations and human perceptions. Results from Study 1 showed the AI models consistently portrayed men as having the greatest desire for social distance from someone with mental health problems, relatively higher Stereotypes scores, and the least favorable attitudes toward help-seeking, compared with women and nonbinary individuals. Study 2 confirmed that this ordinal pattern partially aligned with human-reported data for Personal and Self Stigma. However, the AI models systematically underestimated Personal Stigma and Perceived Stigma while substantially overestimating Self Stigma for all gender groups. These findings reveal that, although the AI models have learned relative gender-based patterns reflecting human experiences to a certain extent, they function as distorted portrayals, highlighting the need for rigorous, context-specific validation against real-world data to ensure AI's safe and equitable deployment.