Gender cues influence ADHD self-diagnosis responses in GPT models
Gender Associations in LLM-Mediated ADHD Self-Diagnosis
Computation and LanguageComputers and Society
Summary
Large language models like GPT are used by people to help self-diagnose medical conditions such as ADHD. This study looked at how gender hints given to the model affect its ADHD diagnosis answers. The authors found that explicitly telling the model about gender changes the kind of ADHD it associates with more than just implying it indirectly. These findings show that the model’s responses can include biases tied to gender, and the way these cues are given matters.
What this means in practice
- •For healthcare chatbot developers: Design chatbots that consider how explicit gender cues affect ADHD-related diagnostic suggestions from language models.
- •For clinical ai safety teams: Evaluate bias mitigation strategies in AI tools by analyzing how demographic cues alter diagnostic outputs for conditions with known gender disparities.
Authors
Matilde Elene Hansen, Ross Deans Kristensen-McLachlan
Abstract
Large Language Models (LLMs) are increasingly used for medical self-diagnosis. However, the way demographic cues shape outputs in these interactions remains poorly understood. This is particularly consequential for diagnoses like ADHD, where persistent gender disparities are not fully aligned with the clinical evidence. We adapt the Context Association Test (CAT) to evaluate whether GPT-4o responds to gendered ADHD associations in a zero-shot, self-diagnostic setting. We show that explicit cueing consistently amplifies certain ADHD subtype associations while implicit cueing produces near proxy-aligned distributions. These divergences suggest that explicit demographic cues may activate bias-mitigation mechanisms that over-correct rather than eliminate bias, whereas implicit cues appear to reflect the model's default behavior. Our results demonstrate that evaluating gendered bias in LLM-mediated self-diagnosis requires careful attention to the demographic associations a model encodes.