Review proposes clinical model for AI decision oversight

Researchers have proposed a new conceptual framework designed to address the ethical gaps in artificial intelligence-mediated clinical decision-making. The study, published in BMC Medical Ethics, presents a critical narrative review and a specific encounter-level deliberative model intended to help institutions manage the integration of AI in diagnosis, prognosis, triage, treatment, and resource allocation. The authors argue that while existing ethical and governance frameworks identify important values and lifecycle duties, healthcare institutions still lack a shared procedure that specifies who must act, what must be checked and documented, and when an AI-supported decision should be overridden, paused, or escalated.

The primary objective of the work was to map recurrent ethical and governance requirements within a purposively assembled source corpus. The researchers used this mapping to construct an encounter-level deliberative model, explicitly noting that the model is a conceptual proposal rather than an automatic finding derived from the literature.

The methodology involved a critical narrative review with conceptual mapping of a heterogeneous source set identified through iterative scoping between January and March 2025. The review is explicitly described as not systematic, exhaustive, or representative, a limitation the authors acknowledge to maintain transparency. The master set initially contained 54 records, from which 11 duplicates and one non-evidence record were removed, leaving 42 retained sources. Of these, 35 informed the primary conceptual mapping, while seven served foundational, methodological, post-search, or companion-publication roles. Three members of the review team participated in source selection, and the preserved records do not support a claim of independent triple screening or a retrospective agreement statistic.

From this source corpus, the researchers mapped five recurrent domains that are central to the ethical use of AI in healthcare. These domains include accountable clinical authority and meaningful human oversight; clinically meaningful explanation, uncertainty communication, and contestability; equity, bias, and distributive impact; lifecycle safety, security, and accountability; and proportional data governance, privacy, and patient agency. These domains served as the foundation for the author-developed Deliberative Clinical Cyberbioethics Model, or D-CCM.

The proposed D-CCM is a seven-step model that specifies accountable actors, mandatory inputs and records, and four possible outputs for each decision point: continue, continue with safeguards, pause or stop, or escalate. The model also includes explicit trade-off rules and identifies four forms of drift or divergence that can occur in AI-mediated processes. Each step also specifies escalation triggers.

The model further incorporates five candidate patient-facing protections and mechanisms for scheduled and trigger-based revision. Comparison with WHO, UNESCO, NIST AI RMF, and FUTURE-AI positions the D-CCM as a complementary point-of-care translation layer.

The authors conclude that Clinical Cyberbioethics is presented as a proposed integrative governance lens rather than an established new discipline. The D-CCM is a practice-oriented conceptual proposal that requires prospective co-design, feasibility testing, reliability assessment, and outcome evaluation before it can be treated as a validated operational framework.

Subject of Research: A conceptual proposal for an encounter-level deliberative model in clinical cyberbioethics for AI-mediated decision-making.

Article Title: Clinical cyberbioethics for AI-mediated clinical decision-making: a critical narrative review and conceptual proposal for an encounter-level deliberative model

Article References: Díaz-Pérez, A., & Pérez, W. J. A. (2026). Clinical cyberbioethics for AI-mediated clinical decision-making: a critical narrative review and conceptual proposal for an encounter-level deliberative model. BMC Medical Ethics. https://doi.org/10.1186/s12910-026-01596-z

Image Credits: AI Generated

DOI: 10.1186/s12910-026-01596-z

Keywords: clinical cyberbioethics, AI-mediated decision-making, deliberative model, narrative review, human oversight, clinical authority, uncertainty communication, equity and bias, lifecycle safety, data governance, patient agency, D-CCM model, WHO guidelines, UNESCO frameworks, NIST AI RMF, FUTURE-AI, point-of-care translation, ethical governance, resource allocation, clinical trials