New method advances efforts to overcome bias in AI tool for children with anxiety

Researchers in Cincinnati and London reduced AI bias in pediatric mental health diagnostics by optimizing training data. Their method cut gender-based performance gaps by 27% across 20,000 cases.