AI-natomy: human anatomy through the eyes of artificial intelligence. Is there a distinction between reality and imagination?
Year: 2025
Type: article
Abstract: BACKGROUND: The application of artificial intelligence (AI) in medicine has increased markedly in recent years. Many generative tools use a model that processes a text command (a so-called "prompt") and generates a response in the form of text (e.g., ChatGPT) or an image (e.g., Midjourney, DALL-E). Because of their wide availability, these tools are often used for educational purposes without appropriate supervision. This study aimed to assess the current state of AI-based generative tools for creating an image-based anatomical atlas using only AI-generated images. MATERIALS AND METHODS: The authors entered anatomical prompts into Midjourney, and individual images were created for each of the eight organ systems: digestive, facial, genitourinary, respiratory, skeletal, muscular, circulatory, and neurological. The systems were divided into two sections: a general sketch of the body system (BS) and a detailed description of one representative structure (DS). The authors selected the two highest-quality images for each collection. The obtained data were assessed qualitatively according to the presence of a specific structure, physiological arrangement, and the presence of additional AI hallucinations. The obtained results were compared with the Federative International Programme for Anatomical Terminology, Terminologia Anatomica 2nd version (TA2), and statistically analyzed using the chi-square test of independence (two-sided α = 0.05, β = 0.2) to evaluate the relationship between image complexity and the studied categories. RESULTS: A total of 12,126 general BS items and 2,503 specific DS items according to TA2 were obtained and analyzed. Regarding the presence of structures, they were generally present in 17.3% (n = 2,094), absent in 53.3% (n = 6,462), and the generated scenery did not allow a clear assessment of structural presence in 29.4% (n = 3,570). Only 7.0% (n = 553) of the present structures were in physiological order. However, 6.0% (n = 471) of the specific items were classified as AI hallucinations. When comparing BS and DS findings, the tool performed significantly better for DS in all categories: structural presence (21.7% BS vs 40.3% DS; p < 0.001), physiological order (4.6% BS vs 15.5% DS; p < 0.001), and AI imagination (6.3% BS vs 4.9% DS; p = 0.034). CONCLUSIONS: The current state of generative tools in anatomy does not allow the high-quality use of AI to generate anatomical atlases. In particular, inexperienced users may unknowingly cultivate inconsistent scientific understanding. The results are expected to improve in the future because of the continuous addition of annotations to libraries and improvements in algorithms; nevertheless, further research is required.
Source: Folia Morphologica
DOI: 10.5603/fm.107607
Language: English
Subfield: Health Informatics
Field: Medicine
Domain: Health Sciences
Open Access status: diamond
OpenAlex ID: W4416358245