Gender Disparities in AI-Driven Depression Detection: A Systematic Review of Algorithmic Bias and Its Implications for Women’s Health
by Anthony, Clement Ogbeh, Dumebi Okuagu, Ihuoma Goodness Dike, Ndorenyin Saviour Udofia, Sandra Ada Collins
Published: April 29, 2026 • DOI: 10.47772/IJRISS.2026.100400058
Abstract
Background: This systematic review critically synthesises evidence regarding gender disparities in AI-driven depression detection, emphasising algorithmic bias and its ramifications for women's health.
Methods: According to PRISMA guidelines, a systematic search of six databases (PubMed, IEEE Xplore, ACM Digital Library, PsycINFO, Scopus, and Web of Science) found 28 studies that met the criteria and were published between 2015 and 2025
Results: The results show that there is a significant disparity in performance between men and women, with models often being less sensitive to depression in women. Bias sources include underrepresentation of female subjects in training data, reliance on male-normative symptom presentation, and feature selection that neglects psychosocial determinants of women's mental health.
Conclusion: The review concludes that current AI models risk perpetuating diagnostic inequities, necessitating the development of gender-inclusive datasets and fairness-aware algorithms.