Patients over Process: Stratifying Risk in the Design, Development, and Deployment of Artificial Intelligence in Healthcare
Bibliographic record
Abstract
The global focus on artificial intelligence (AI) in healthcare and medicine is on the rise. Despite remarkable progress in integrating AI into clinical workflows, gaps in regulation remain a prevalent issue within healthcare systems. Effective regulation of artificial intelligence in clinical practice is essential for managing medico-legal risk and ensuring patient safety. Numerous studies highlight the significant potential for medico-legal risk and the need for clear guidelines on the ethical and safe use of AI in clinical practice. Although there are various concerns that these guidelines must address, our work focused on researching best practices regarding patient-centered factors like patient autonomy, trust and transparency, privacy and security, equity and fairness, and ensuring human oversight. While challenges in AI workflow integration arise from many factors, including human interactions and system inadequacies, the focus on individuals rather than the system has fostered an unsuitable culture for enhancing patient-centered care. Key focus areas include risk stratification strategies and increasing transparency within this inherently complex system, as they play a crucial role in guiding clinical decisions in patient management. Proper integration of AI regulatory frameworks into clinical practice is essential for addressing gaps in the design, development, deployment, and long-term monitoring of AI solutions. Globally, the regulation of AI in clinical practice is continually evolving as governments and legal systems adapt to the rapid advances in AI as a medical device (AIaMD). In Canada, a strategic path forward prioritizes federal and provincial regulations; however, at this stage, they remain fragmented. We advocate for the establishment of uniform guidelines that address the risks, benefits, opportunities, and best practices as AI technologies are integrated into the clinical workflow. Achieving a national standard with clear guidance on the ethical and safe use of AI in clinical practice is recommended to move forward.
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 | 0.117 | 0.188 |
| Meta-epidemiology (narrow) | 0.001 | 0.002 |
| Meta-epidemiology (broad) | 0.001 | 0.002 |
| Bibliometrics | 0.004 | 0.002 |
| Science and technology studies | 0.006 | 0.030 |
| Scholarly communication | 0.026 | 0.032 |
| Open science | 0.003 | 0.017 |
| Research integrity | 0.008 | 0.011 |
| 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".