Lines of Thought / Topic
Clinical AI: the line so far
15 developments and 21 rules, in date order. Built automatically from everything tagged with this topic.
- Rule · In forceINICMR AI health ethics guidelines
- Rule · ProposedIDDraft AI Perpres
- Rule · In forceUS-UTUtah AI Policy Act
- Rule · Partly in forceEUEU AI Act
Also on Project Meridian and the race to govern military AI, Transparency, not licensing, Who decides if a medical AI is safe?
- Rule · In forceUS-UTUtah mental-health chatbot law
- Rule · In forceKEKenya National AI Strategy
- Rule · In forceUSOMB M-25-21
- FDA rolls out 'Elsa', a generative AI assistant for all agency staff
The regulator that clears medical AI is now using generative AI inside its own review work, raising questions about validation and accountability for regulatory decisions.
- Illinois bars AI from delivering therapy without a licensed professional
Mental-health chatbot makers now face a patchwork in which the same product is banned in some US states and merely labelled in others.
- Rule · In forceUS-ILIllinois AI therapy ban (WOPR Act)
- Rule · Partly in forceUS-CACCPA ADMT regulations
Also on Can an algorithm deny your care?, Who audits the hiring algorithm?
- Rule · In forceITItalian AI Law
- Rule · In forceUS-NYNY AI companion law
- FDA gives all employees agentic AI tools after 70% take-up of Elsa
Agentic tools inside premarket review and inspections could speed approvals but make it harder for outsiders to see how regulatory judgements are reached.
- Rule · In forceUSState AI law preemption EO
Also on Transparency, not licensing, Can an algorithm deny your care?
- FDA loosens oversight of AI decision-support software and health wearables
More AI tools used by clinicians and consumers can reach market without FDA clearance, shifting safety assurance to developers and health systems.
- OpenAI launches ChatGPT Health, linking medical records and wellness apps
A mass-market chatbot now ingests patient records while positioning itself outside medical-device rules, testing where consumer health information ends and regulated advice begins.
- Anthropic launches Claude for Healthcare with HIPAA-ready tools for providers and payers
Frontier labs are now competing directly for payer and provider workflows such as prior authorisation, where errors translate into denied or delayed care.
Also on Can an algorithm deny your care?
- Google releases MedGemma 1.5 and MedASR open medical models
Open-weight medical models let hospitals and startups, including in India, build imaging and scribe tools without relying on closed APIs, while leaving validation to each deployer.
- MASAI trial: AI-supported mammography screening cuts interval cancers by 12%
It is the first randomised evidence that AI in population screening improves a hard clinical outcome, strengthening the case for national programmes to adopt it.
- Rule · ProposedKEKenya AI Bill
- Washington law says only a licensed clinician can deny prior-authorisation requests
Insurers and AI vendors must build human sign-off and audit trails into utilisation-review tools, state by state.
Also on Can an algorithm deny your care?
- NHS England updates AI scribe guidance and launches supplier registry
It turns AI scribes from local experiments into a regulated NHS product category with a minimum compliance bar for suppliers.
- Science study: OpenAI's o1 beat physicians at diagnosis on real ER cases
Strong retrospective results increase pressure to deploy diagnostic LLMs, which makes prospective trials and clear regulatory pathways more urgent.
- EU deal keeps AI medical devices under AI Act high-risk rules, delayed to 2028
Medtech makers selling AI devices in Europe must plan for dual MDR and AI Act conformity, with one extra year to prepare.
- Rule · Enacted, not yet in forceUS-COSB 26-189 (Colorado ADMT law)
- Rule · In consultationEUHigh-risk classification guidelines (draft)
- Rule · In forceEUDigital Omnibus on AI
- FDA proposes a risk framework for generative AI medical devices
It is the first concrete sign of how FDA might clear LLM-based clinical tools, and comments will shape whether chatbots and scribes need premarket review.
- DeepMind releases AlphaGenome Atlas scoring all 9bn single-letter human DNA changes
Free, precomputed variant-effect predictions lower the barrier for clinical genetics and rare-disease research worldwide.
- Rule · Enacted, not yet in forceUS-CASB 1119 / Adam's Law
- UK accepts all 44 recommendations of doctor-led commission on regulating health AI
Medical-AI makers selling in the UK should expect ongoing monitoring duties after approval, not a one-off clearance, bringing the UK closer to the FDA's lifecycle approach.