A Study on the Utility of Artificial Intelligence in the Draw-A-Story Assessment

This study examined the applicability of AI-assisted scoring in the Draw-A-Story (DAS) assessment. Ninety-nine adults completed the DAS and the Beck Depression Inventory. Two certified art therapists independently scored drawings across four domains—emotional content, self-image, color usage, and compositional transformation—and ratings were compared with GPT-4–generated scores. Significant group differences emerged in emotional content and self-image by depression level. Quadratic weighted kappa indicated substantial agreement between AI and experts (κ = .67, p < .001), with higher concordance in color usage than in symbolically interpreted domains. Art therapists considered AI-generated descriptions clinically appropriate and useful supplementary information. Findings suggest AI may enhance scoring consistency while supporting, not replacing, clinical judgment.