Does any of this work?¶
A governance tool should be honest about its own evidence base. This page summarizes what is actually known about whether AI governance frameworks change anything, and what remains unmeasured.
External recognition¶
In June 2026 the WHO Regional Office for Europe published Bridging theory and practice — implementation insights on artificial intelligence in health care, the report of Working Group 1 of the Strategic Partners' Initiative for Data and Digital Health (WHO/EURO:2026-13225-52999-82657).
The report cites a comprehensive review that identified more than a dozen reporting and evaluation frameworks, and then presents four complementary approaches. Two of them are the two frameworks this tool implements:
- OPTICA, described as a practical evaluation tool for assessing AI solutions toward adoption, 77 items in 13 chapters and four domains, requiring input from five designated stakeholders.
- CHAI, described as a tool for AI solution transparency, whose applied model card acts as a standardized "nutrition label" for an AI solution.
The report notes (printed page 8, PDF page 14) that OPTICA "embeds requirements for organizational AI governance, which include continuing monitoring and impact assessment." It describes these approaches. It does not endorse any of them.
Why the two are on this site
WHO presents the four approaches as complementary. This project's own premise is that OPTICA and CHAI complement each other. That is the premise of the crosswalk: OPTICA supplies the adoption-time depth, CHAI supplies the lifecycle and the transparency artifact. Neither subsumes the other.
What the evidence actually shows¶
A 2025 scoping review in JAMIA — Chan, Rahimi-Ardabilli, Rogers and Coiera, The real-world impact of artificial intelligence ethics frameworks across a decade in healthcare — looked for studies of frameworks that had actually been put into practice.
The headline numbers are sobering:
| Published AI ethics frameworks identified | at least 173 |
| Studies of frameworks actually operationalized | 16 |
Of those 16, the review found three ways frameworks get used: establishing governance structures (4), pre-implementation ethical review (8), and post-deployment audit (8). The categories overlap, so the counts add up to more than 16.
OPTICA appears among them. Applied to 18 AI solutions at Clalit, it was reported to produce "improved oversight and accountability for AI use" and "more informed decisions on AI solution deployment against predefined metrics."
The gap you should know about¶
No framework has been shown to improve patient outcomes
The review's central finding is blunt:
No studies demonstrated a direct link between AIEFs and health-related outcome measures such as patient safety.
It also notes that no study reported health-related outcomes such as increased quality of care, and that all 16 used cross-sectional designs with no before-and-after comparison.
What that means in practice: completing a checklist in this tool is evidence that a governance process happened and was documented. It is not evidence that patients are safer. The honest claim for any tool in this category — including this one — is that it makes decisions legible, attributable and reviewable. The link from that to clinical benefit is assumed, not demonstrated.
The review's own recommendation is to measure the intermediate stages rather than claim the endpoint. That is a reasonable thing to ask of your own governance program too: track whether reviews changed decisions, not merely whether they were completed.
Sources¶
- WHO Regional Office for Europe (2026). Bridging theory and practice — implementation insights on artificial intelligence in health care: report of Working Group 1 of the Strategic Partners' Initiative for Data and Digital Health. WHO/EURO:2026-13225-52999-82657
- Chan A, Rahimi-Ardabilli H, Rogers WA, Coiera E (2025). The real-world impact of artificial intelligence ethics frameworks across a decade in healthcare: a scoping review. J Am Med Inform Assoc. PMC12626214
- Dagan N, Devons-Sberro S, Paz Z, et al. (2024). Evaluation of AI Solutions in Health Care Organizations — The OPTICA Tool. NEJM AI 1(9). DOI: 10.1056/AIcs2300269