Evaluation of artificial intelligence clinical applications: Detailed case analyses show value of healthcare ethics approach in identifying patient care issues
Why this work is in the frame
A frame that forgets how it found something cannot be audited. These are the routes that admitted this work.
Bibliographic record
Abstract
Abstract This paper is one of the first to analyse the ethical implications of specific healthcare artificial intelligence (AI) applications, and the first to provide a detailed analysis of AI‐based systems for clinical decision support. AI is increasingly being deployed across multiple domains. In response, a plethora of ethical guidelines and principles for general AI use have been published, with some convergence about which ethical concepts are relevant to this new technology. However, few of these frameworks are healthcare‐specific, and there has been limited examination of actual AI applications in healthcare. Our ethical evaluation identifies context‐ and case‐specific healthcare ethical issues for two applications, and investigates the extent to which the general ethical principles for AI‐assisted healthcare expressed in existing frameworks capture what is most ethically relevant from the perspective of healthcare ethics. We provide a detailed description and analysis of two AI‐based systems for clinical decision support (Painchek ® and IDx‐DR). Our results identify ethical challenges associated with potentially deceptive promissory claims, lack of patient and public involvement in healthcare AI development and deployment, and lack of attention to the impact of AIs on healthcare relationships. Our analysis also highlights the close connection between evaluation and technical development and reporting. Critical appraisal frameworks for healthcare AIs should include explicit ethical evaluation with benchmarks. However, each application will require scrutiny across the AI life‐cycle to identify ethical issues specific to healthcare. This level of analysis requires more attention to detail than is suggested by current ethical guidance or frameworks.
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.
Full frame distilled prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. Learned from the 10,348 direct Codex labels and 10,348 direct Gemma labels. Candidate is the union of thresholded teacher heads; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels or direct frontier model labels.
Codex and Gemma teacher scores by category
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.006 | 0.005 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.001 | 0.000 |
| Bibliometrics | 0.000 | 0.001 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.000 | 0.000 |
| Open science | 0.000 | 0.000 |
| Research integrity | 0.001 | 0.001 |
| Insufficient payload (model declined to judge) | 0.000 | 0.000 |
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 it