Who published AI guidance for health, and when
Only AI-specific guidance is shown, newest first, back to the first major AI-ethics frameworks in 2019. Each entry shows the month it was first published, who issued it and what kind of document it is. Choose a topic to follow one thread.
Expected next: 16 announced documents and dates ↓
Dates are the first publication of the version we link to (online date for journal articles). A dash means we have not yet confirmed the month.
Expected next
AI guidance that has been announced but is not yet published or in force. We list an expected date only when the developer or the authority has stated one, and mark items that have missed their own date as delayed. Checked 1 Nov 2026.
- November 2026
A regulatory sandbox that tests how AI medical devices are monitored once they are in use.
First cohort of a three-year phase; applications open.
- Q4 2026 (target)
Technical standards on risk management, data, logging, human oversight, accuracy and cybersecurity that high-risk AI can follow to show it meets the AI Act.
In development; only the quality-management standard (EN 18286) is published so far.
- December 2026
How AI devices that are updated or retrained after approval should manage and report those changes.
Promised in the UK government’s response to the National Commission.
- Consultation closes December 2026
Professional standards for safe use of AI by nurses and midwives in the UK.
Public consultation September to December 2026. The new Code is expected to include duties on AI and digital technology.
- Early 2027
When AI software counts as a medical device in the UK, and which risk class it falls into.
Announced.
- Spring 2027
A full plan for regulating health AI across its life cycle in the UK.
Announced; will set out how the National Commission’s 44 recommendations are put into practice.
- August 2027
Practical rules for deciding whether an AI system is high-risk under the AI Act.
Required by the 2026 amendment of the AI Act.
- By September 2027
What makers must submit for AI medical devices, and how to manage them over their life cycle.
Draft since January 2025; on the FDA’s top-priority list for final guidance in fiscal year 2027.
- By September 2027 (draft)
The evidence needed for conversational, generative-AI devices that address mental disorders.
On the FDA’s fiscal year 2027 priority list.
- 2 December 2027
Risk management, data quality, human oversight and documentation duties for listed high-risk uses, such as access to essential services, education and employment.
Adopted; postponed to this date by the 2026 amendment.
- 2 August 2028
The same high-risk duties for AI that is part of products such as medical devices and in-vitro diagnostics.
Adopted; postponed to this date by the 2026 amendment.
- No date announced
The first binding international treaty on AI and human rights.
Signed in 2024; the EU ratified it in May 2026. It enters into force three months after five ratifications, including three Council of Europe member states.
- Was due December 2025COREQ+LLMDelayed
Additions to the COREQ checklist for qualitative studies that use large language models.
Protocol published 2025; expert consensus (Delphi) under way.
- Was due early 2025QUADAS-AIDelayed
A tool for judging risk of bias in studies of the diagnostic accuracy of AI.
Protocol published 2024; no final tool yet.
- Was due 2022PRISMA-AIDelayed
Reporting items for systematic reviews and meta-analyses of AI studies in health.
Registered and announced in 2023; no final guideline yet.
- No date announcedCANGARUDelayed
Standards for disclosing and reporting the use of generative AI in academic writing.
Protocol published 2023; no final guideline yet.