Artificial Intelligence in Action: Racial and Gender Disparities in Academic Radiology
Bibliographic record
Abstract
Academic radiology continues to face persistent gender and racial disparities in career advancement. The emergence of generative artificial intelligence (AI) platforms offers new opportunities to analyze workforce diversity patterns rapidly and at scale. This study aimed to evaluate the interpretative capacity of three generative AI platforms (i.e., ChatGPT, DeepSeek, and Perplexity) in identifying disparities in academic rank and tenure status across gender and racial subgroups in academic radiology. The outputs of these AI models were compared with conventional human-led analyses for accuracy, limitations, and potential biases. We prompted each AI model to analyze publicly available American Association of Medical Colleges Faculty Roster data on tenure and rank distribution by gender and race using standardized query templates. Outputs were systematically evaluated for consistency, accuracy, and potential biases against benchmark human-curated studies. Comparative analysis included variations between AI platforms and traditional research methods, with particular attention to how each model interpreted and reported disparities. The AI models broadly recognized trends in faculty growth and underrepresentation, but interpretations varied. Perplexity and DeepSeek provided more granular insights, such as declining tenure rates and intersectional disparities, while ChatGPT offered less detailed analyses. Key discrepancies included divergent temporal trends and policy recommendations, highlighting AI's limitations in capturing nuanced sociodemographic patterns. Generative AI shows promise in analyzing workforce disparities but requires validation to mitigate biases and inconsistencies. When used alongside traditional methods, AI can enhance understanding of inequities in academic medicine, provided that its outputs are critically evaluated for fairness and accuracy.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.024 | 0.102 |
| Meta-epidemiology (narrow) | 0.001 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.001 |
| Bibliometrics | 0.004 | 0.004 |
| Science and technology studies | 0.002 | 0.003 |
| Scholarly communication | 0.006 | 0.005 |
| Open science | 0.002 | 0.004 |
| Research integrity | 0.001 | 0.002 |
| Insufficient payload (model declined to judge) | 0.004 | 0.001 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".