Exploring the Ethical Challenges in the Design and Auditing of a Machine Learning Computer Vision Algorithm for Clinical Use
Bibliographic record
Abstract
Faulty prediction models and hidden racial biases have been found in algorithms affecting the care of hundreds of millions of patients, bringing the apparent risks of the integration of artificial intelligence into healthcare to public attention. This retrospective case study describes the most ethically salient issues encountered by those involved in the design of a deep learning-powered computer vision radiology algorithm intended for use in clinical care. The participants of this study perceived that the cultures and expectations associated with professional academic norms in AI and adjacent disciplines tended to promote the goals of AI innovation over the needs of health systems, or specific clinical use cases. They identified several ways in which more prescriptive regulatory requirements could better meet the needs of health systems. When participants situated the design process within its social, economic, or historical contexts, they took a more reflexive approach to the ethical issues they encountered.
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 | 0.089 | 0.180 |
| Meta-epidemiology (narrow) | 0.000 | 0.001 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.001 | 0.001 |
| Science and technology studies | 0.006 | 0.011 |
| Scholarly communication | 0.010 | 0.006 |
| Open science | 0.002 | 0.004 |
| Research integrity | 0.004 | 0.007 |
| Insufficient payload (model declined to judge) | 0.002 | 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".