Evaluation of Image Analysis Using General-Purpose AI to Identify Causes of Residential Fall Accidents and Suggest Improvements

At present, artificial intelligence (AI) has demonstrated remarkable advances in multimodal AI. In this study, we employed a widely used, general-purpose AI without additional training to extract spatial information from multiple indoor photographs taken at the sites of fall accidents in residential environments. The AI was then tasked with analyzing the causes of the falls and suggesting countermeasures. The results indicated that spatial recognition itself was difficult in restrooms and bathrooms. However, in other rooms, the AI generally provided appropriate responses regarding spatial understanding, explanations of fall circumstances, and recommendations for improvements.