AI must be licensed as a doctor
„Generative AI must be qualified based on competence, similar to doctors, and can perform clinical tasks without real-time medical supervision if appropriate safeguards are in place.”
Counter-narrative
The Federation of State Medical Boards (FSMB) states that generative medical AI should not be licensed as an independent 'practitioner.' Nurses protest against AI care without doctors, and the legal profession (Law360) states that the competence-based approach raises legal concerns. California has codified into law that medicine is practiced by doctors and nurses, not AI.
Carriers
FDA CDRH, Vinod Khosla, Utah's regulatory sandbox (Nolla Health), drafters of the Senate bill, Vara and autonomous AI suppliers, Haider Warraich
Analysis
The narrative represents the boldest regulatory shift: moving from a device-based view to a classification resembling a practitioner. It is fueled by the FDA (discussion paper, TEMPO pilot), an investor (Khosla requesting certification for AI surpassing the 'average doctor'), and the Trump administration's health department, whose pilot, according to Warraich, aims at long-term deployment of autonomous AI. California's countermove, stating that medicine is practiced by doctors and nurses, shows the rift between states. The FDA and CMS are trying to avoid regulatory congestion due to state AI clinical trials. The weakness is that the chain of responsibility is not settled. Professional reactions (ATA) highlight manufacturer liability, while the competency-based approach raises legal concerns.
What's at stake: Whether a human remains responsible for the clinical decision, or certified software takes over.
Narrative sub-threads
FDA's competency-based generative AI framework
In a discussion paper, the FDA proposes that generative AI tools be qualified on a competency basis similar to physicians: based on benchmarks, then clinical confirmation, and finally evaluation of real-world use. This is accompanied by pilots: one offering a pre-approval path to patients, while the TEMPO pilot tests generative AI tools.
This strand is driven by the federal device regulator and organized around the comment (consultation) process. Regulatory planning and pilots are at the center, not mature, broad clinical deployment.
↗ rising · evidence: strongState sandboxes: prescribing without doctors
In regulatory sandboxes, AI already performs clinical tasks without direct medical supervision. In Utah, the Nolla Health platform evaluates acne based on photos and can issue initial prescriptions, currently only for acne treatment. The first hundred patient prescriptions must be pre-approved by two doctors.
Here, the focus is not on a draft but on an operational, albeit narrowly scoped, permit. The framework is provided not by a federal but by a state or national sandbox. This is supplemented by FDA and CMS efforts to avoid regulatory bottlenecks.
↘ fading · evidence: mediumCertifying autonomous diagnostic AI
Autonomous decision-making AI has appeared in imaging screening. Vara received CE marking for autonomous breast screening, and some mammograms may conclude without radiologist review, though clinical benefit remains uncertain according to experts. ARPA-H funding supports the development of the first autonomous cardiac AI at Tempus, up to $9.5 million.
This branch is carried by diagnostic products and screening practice, not by the medical-licensing analogy. The evidence consists of product authorisation, research funding and a screening study; the latter measures AI-assisted, not fully autonomous, reading.
↗ rising · evidence: moderatePolitical and funding pressure for autonomy
Investors, government actors, and Senate bills are pushing the system toward greater clinical autonomy for AI. Khosla is asking the FDA to certify autonomous AI that outperforms the average doctor. Kennedy stated that AI can provide a better-informed second opinion than doctors. A Senate bill suggests Medicare could experimentally pay for clinical AI operating without real-time medical review.
The actors in this strand are political and capital-side voices, not the device-approval procedure. The mechanism is reimbursement and the political agenda; the evidence is statements and an outline of a draft bill. The claims were criticized by doctors and nurses.
→ lasting · evidence: moderateThe profession defends the human license
The Federation of State Medical Boards says generative medical AI should not be licensed as an independent practitioner. Oregon, Washington and Delaware prohibit software from using the title of nurse. California has written into law that medicine is practiced by doctors and nurses, not AI. Newsom, however, vetoed a bill protecting nurses who override AI recommendations.
This is the counter-narrative: professional self-regulatory bodies and state legislation do not extend the medical analogy but restrict it. The closure is not complete, however, as the California veto shows.
Sub-thread stories
FDA's competency-based generative AI framework
FDA discussion paper would assess generative AI devices in a way similar to doctorsmedtech companyDuring the comment period, companies should indicate which tier of the proposed graduated assessment their product would fall into and what evidence they can provide for it.doctorThrough professional organizations, physicians can have a say in what clinical confirmation generative AI tools should be required to show.health ITTeams need to assess which embedded generative functions could fall under device regulation.health policyThe FDA is signaling a direction for authorizing generative AI, which may serve as a benchmark for other regulators.FDA publishes discussion paper on regulating generative AI-enabled medical devicesmedtech companyCompanies should comment on the paper, because the planned guidance will set the authorization pathway for generative AI products.health ITTeams developing or procuring software with generative functions should anticipate future FDA expectations.health policyThe public consultation offers a model for regulating fast-changing generative AI within device law.doctorThe status and evidence base of generative tools used on the client side may become clearer as a result of the regulation.Legal concerns raised over FDA's 'competency' approach to AI regulationmedtech companyGiven the legal uncertainty, compliance strategy and regulatory communication should be reviewed from a legal perspective too.health policyThe legal basis of the regulatory approach may be contested, which could affect enforceability.FDA pilot offers generative AI medical devices a path to patients before authorizationmedtech companyCompanies developing generative AI should examine whether to take part in the pilot and prepare for post-market monitoring.health policyThe regulatory approach could offer a model for other authorities handling generative AI devices.doctorUnder the pilot, devices may reach patients before authorization, so clinicians need to know a device's validation status.FDA weighs regulating AI devices like doctors as TEMPO pilot tests generative AImedtech companyThe approval pathway for generative AI devices is taking shape in the TEMPO pilot; it is worth following the participation criteria.doctorIf devices are regulated by a doctor-like standard, questions of liability and oversight may change.health policyThe FDA's direction could set a precedent for other regulators in handling autonomous AI.FDA publishes discussion paper on regulating generative AI-enabled medical devicesmedtech companyThe discussion paper is an opportunity to give feedback, so it is worth submitting expert comments.health policyDevelopments in the regulatory framework for generative AI devices should be followed.FDA extends feedback period on generative AI medical device rules and tests language models on radiology reportsmedtech companyThe regulatory framework for generative AI devices is still taking shape, so companies should use the opportunity to comment.health policyThe FDA's experiments point to future methods of evaluating AI.FDA explores competency-based path for generative AI-enabled medical devicesmedtech companyShould comment on the framework and prepare for stricter evidence of intended use and clinical risk.health ITMust align generative AI CDS development with the lifecycle-based approach.health policyShould follow how the competency-based approach, modelled on the way physicians are assessed, develops.FDA outlines 2027 AI guidance plans; nurses can comment on generative AI framework until October 19medtech companyDevelopers of generative AI devices need to follow the FDA framework, as it will shape future authorisation expectations.health policyThe FDA's 2027 plans point to the direction of future oversight of AI devices.FDA publishes discussion paper on generative AI-enabled medical devicesmedtech companyCompanies should submit comments before the deadline, as the discussion paper lays the groundwork for the future regulatory framework for generative AI devices.health policyThe planned framework could also become a benchmark for regulators in other markets.AHA, AMA, ATA and radiology groups respond to FDA plan for generative AI medical devicesmedtech companyRisk-based classification, PCCPs and manufacturer accountability bear directly on the authorization plans for generative AI products.doctorThe professional bodies' proposals may also shape how clinician responsibility and oversight are divided.hospital managerDiverging FDA and state rules create procurement and compliance risk for generative AI tools.health policyThe conflict between state and federal levels creates an unclear regulatory environment.Stakeholders offer competing views on how FDA should regulate generative AI in medtechmedtech companyApproval strategies for generative AI products must be adjusted as the FDA's approach evolves.health policyIn shaping the regulatory framework, the challenge is balancing the differing interests of stakeholders.
State sandboxes: prescribing without doctors
Israel's regulatory sandbox tests autonomous AI in clinical carehealth policyThe sandbox is a model for gaining regulatory experience in a controlled setting.medtech companyIt offers a testing opportunity for autonomous AI tools.Utah lets AI prescribe acne medication without direct physician oversightdoctorTwo physicians' approval is required for the first 100 prescriptions, so the supervisory role remains for now, but part of routine prescribing could shift to AI.health policyUtah's model is a template for the regulatory sandbox and AI prescribing, worth watching because of its relationship with the FDA.hospital managerIt is worth assessing how the reshaping of routine care affects outpatient volumes and liability questions.health ITIntegrating photo-based assessment and prescription issuing raises new vendor and data-handling requirements.Utah approves AI system that prescribes medication without physician oversightdoctorThis is a precedent for handing the prescribing role to AI, so professional chambers will need to take a position on liability.health policyThe model is worth watching as a possible template or a cautionary example in regulating prescribing authority.pharmaceuticalsDirect AI prescribing could also change the access path for prescription-only products.FDA and CMS scramble to avoid regulatory logjam as more states look to test AI in clinical settingshealth policyMisaligned state and federal rules could create uncertainty for AI pilots.hospital managerLicensing and payment conditions should be clarified before joining state AI experiments.AI writes acne prescriptions without a doctor's visit in one US statedoctorAutonomous AI prescribing raises new questions of professional liability and patient safety.health policyThe rules on AI's authority to prescribe need to be clarified.
Certifying autonomous diagnostic AI
Vara receives world-first CE certification for autonomous AI in breast cancer screeningdoctorRadiologists need to prepare for AI working independently in some screening cases.hospital managerBreast screening programs should consider procuring autonomous AI and redesigning their workflows.medtech companyAutonomous diagnostic AI now has an EU regulatory pathway, which is also a benchmark for competitors.AI in breast screening can clear some mammograms without radiologist review, but clinical benefit is uncertaindoctorThe radiologist's role in the screening protocol is changing, so sample-based checks of AI reads are worth setting up.hospital managerWhen planning AI for screening programmes, demand real, local outcome data.medtech companyA standalone clearing function will require stricter evidence from regulators and buyers.ARPA-H backs agentic AI for cardiovascular care, with awards to Atman Health and TempusdoctorAutonomous AI decision support in cardiology is moving closer to clinical practice.medtech companyARPA-H funding gives a competitive edge in autonomous cardiology AI.AI plus one radiologist matches two radiologists in breast cancer screening, analysis findsdoctorReplacing the second reader with AI could mean a change in screening protocols for radiologists.hospital managerReorganising screening with AI could help address radiology capacity shortages.health policyThe study is an argument for reviewing screening guidelines to allow AI-supported reading.
Political and funding pressure for autonomy
Khosla asks FDA to certify AI that outperforms the median doctor; nurses objectdoctorThe regulatory debate over reshaping the physician's role directly concerns the profession's autonomy.US health officials deploy medical AI as safety concerns persisthealth policyThe tension between state-level deployment pilots and federal oversight is a regulatory risk.doctorThe move toward autonomous AI raises questions of liability and oversight.medtech companyNo generative AI product for mental health has yet been authorized, which is both a market gap and a regulatory risk.Kennedy says AI is "better informed" than doctors; physicians push backdoctorDoctors must handle patients' AI-generated "second opinions" and argue professionally for or against them.health policyThe effect of official federal AI endorsement on patient information and trust needs to be monitored.Senate draft would let Medicare test paying for AI that performs clinical tasks without real-time physician reviewdoctorA payment model for AI without physician oversight raises questions of professional liability and division of tasks.health policyThe payment model could be the first major state-funded trial of clinical AI autonomy.medtech companyIt could open a reimbursement path for autonomous clinical AI tools.Senate draft bill seeks input on health AI oversight as FDA and CMS look to work with stateshealth policyCoordinating federal and state regulation is the key issue for the coming period.medtech companyMulti-level oversight may bring differing requirements, so the legislative process must be followed.
The profession defends the human license
Federation of State Medical Boards: generative medical AI should not be licensed to practice on its owndoctorAI should continue to be treated as a tool, with responsibility remaining with the physician.health policyIndividual physician licensing is not the model to follow for regulating AI.Newsom vetoes bill protecting nurses who override AI, but signs clinical judgment lawhospital managerHospital executives must set out in internal policy that clinical judgment takes precedence over AI.health policyCalifornia's twin decision may offer a model for other states' regulation.California law says medicine must be practised by doctors and nurses, not AIdoctorThe law reaffirms that clinical decisions remain the physician's responsibility, even with AI support.hospital managerWhen deploying AI systems, hospitals must ensure that a qualified professional still makes the clinical decision.health policyThe contrast between the Californian and Utah approaches is a reference point for further state regulation.Four US states set different rules on whether AI can call itself a nursemedtech companyProduct names must be checked state by state.California bill aims to ensure AI does not bind doctors' decisionsdoctorThe bill would protect clinical decision-making freedom against AI recommendations.health policyIt could serve as a state-level regulatory model for other states.
Google search interest
Hungary
| Topic and search term | Weekly | Average |
|---|---|---|
| AI must be licensed as a doctormedical license | 0.8 | |
| FDA's competency-based generative AI frameworkFDA regulation | 0 | |
| State sandboxes: prescribing without doctorsacne treatment | 0 | |
| Certifying autonomous diagnostic AImammography | 16.1 | |
| Political and funding pressure for autonomyKennedy healthcare | 0 | |
| The profession defends the human licenseorvosi kamara | 3.9 |
World
| Topic and search term | Weekly | Average |
|---|---|---|
| AI must be licensed as a doctorautonomous ai doctor | 0.1 | |
| FDA's competency-based generative AI frameworkfda generative ai | 8 | |
| State sandboxes: prescribing without doctorsai prescription utah | 0 | |
| Certifying autonomous diagnostic AIai mammography | 1.5 | |
| Political and funding pressure for autonomyarpa-h | 3.1 | |
| The profession defends the human licensemedical licensing board | 49.1 |
Google Trends relative search index over the period studied, weekly average. A narrative and its sub-threads appear in a single comparison: 100 is the group's most-searched day, and the figures can only be compared within the group. The search term was chosen by machine analysis; it measures how much the term is searched, not how widespread the narrative is.
Who is affected and how
- doctor
- The argument in favor of the chamber's position (not an independent practitioner) is strong when you specifically define the supervisory and liability tasks. Take part in the professional consultation: the AMA, the AHA, and radiological societies reacted to the FDA's plan in mid-October.
- medtech company
- The path to autonomy is opening now (FDA pilot, Utah, Vara CE), but so far only for a narrow indication. Plan for a PCCP and real-world data collection, and anticipate greater emphasis on manufacturer liability.
- health policy
- Before opening a sandbox, define the chain of responsibility and shutdown criteria. The FDA–CMS–state gridlock shows what happens without coordination.
Hungary
There is no domestic trace during the examined period. It is worth monitoring how Vara-type autonomous CE markings spread in EU practice.
Weekly trend
Turning points
FDA discussion paper would assess generative AI devices in a way similar to doctorsThe FDA's August 18 discussion paper officially placed a path similar to medical certification on the agenda: benchmarks, then clinical confirmation, and finally evaluation for live use.Vara receives world-first CE certification for autonomous AI in breast cancer screeningOn September 2, Vara became the first in the world to receive a CE certificate for an autonomous breast screening AI. This is a regulatory breakthrough for independently operating diagnostic AI.Utah lets AI prescribe acne medication without direct physician oversightIn a Utah sandbox, Nolla Health's AI can prescribe initial medication without direct medical supervision, currently only for acne. The first hundred patient prescriptions must be pre-approved by two doctors. The debate has moved from theory to practice.Senate draft would let Medicare test paying for AI that performs clinical tasks without real-time physician reviewUnder a Senate bill, Medicare could experimentally pay for clinical AI operating without real-time medical review. This introduces the possibility of reimbursement.
Related narratives in the News magazine
Supporting stories
Federation of State Medical Boards: generative medical AI should not be licensed to practice on its owndoctorAI should continue to be treated as a tool, with responsibility remaining with the physician.health policyIndividual physician licensing is not the model to follow for regulating AI.FDA discussion paper would assess generative AI devices in a way similar to doctorsmedtech companyDuring the comment period, companies should indicate which tier of the proposed graduated assessment their product would fall into and what evidence they can provide for it.doctorThrough professional organizations, physicians can have a say in what clinical confirmation generative AI tools should be required to show.health ITTeams need to assess which embedded generative functions could fall under device regulation.health policyThe FDA is signaling a direction for authorizing generative AI, which may serve as a benchmark for other regulators.FDA publishes discussion paper on regulating generative AI-enabled medical devicesmedtech companyCompanies should comment on the paper, because the planned guidance will set the authorization pathway for generative AI products.health ITTeams developing or procuring software with generative functions should anticipate future FDA expectations.health policyThe public consultation offers a model for regulating fast-changing generative AI within device law.doctorThe status and evidence base of generative tools used on the client side may become clearer as a result of the regulation.Israel's regulatory sandbox tests autonomous AI in clinical carehealth policyThe sandbox is a model for gaining regulatory experience in a controlled setting.medtech companyIt offers a testing opportunity for autonomous AI tools.Legal concerns raised over FDA's 'competency' approach to AI regulationmedtech companyGiven the legal uncertainty, compliance strategy and regulatory communication should be reviewed from a legal perspective too.health policyThe legal basis of the regulatory approach may be contested, which could affect enforceability.FDA pilot offers generative AI medical devices a path to patients before authorizationmedtech companyCompanies developing generative AI should examine whether to take part in the pilot and prepare for post-market monitoring.health policyThe regulatory approach could offer a model for other authorities handling generative AI devices.doctorUnder the pilot, devices may reach patients before authorization, so clinicians need to know a device's validation status.Vara receives world-first CE certification for autonomous AI in breast cancer screeningdoctorRadiologists need to prepare for AI working independently in some screening cases.hospital managerBreast screening programs should consider procuring autonomous AI and redesigning their workflows.medtech companyAutonomous diagnostic AI now has an EU regulatory pathway, which is also a benchmark for competitors.FDA weighs regulating AI devices like doctors as TEMPO pilot tests generative AImedtech companyThe approval pathway for generative AI devices is taking shape in the TEMPO pilot; it is worth following the participation criteria.doctorIf devices are regulated by a doctor-like standard, questions of liability and oversight may change.health policyThe FDA's direction could set a precedent for other regulators in handling autonomous AI.Khosla asks FDA to certify AI that outperforms the median doctor; nurses objectdoctorThe regulatory debate over reshaping the physician's role directly concerns the profession's autonomy.US health officials deploy medical AI as safety concerns persisthealth policyThe tension between state-level deployment pilots and federal oversight is a regulatory risk.doctorThe move toward autonomous AI raises questions of liability and oversight.medtech companyNo generative AI product for mental health has yet been authorized, which is both a market gap and a regulatory risk.FDA explores competency-based path for generative AI-enabled medical devicesmedtech companyShould comment on the framework and prepare for stricter evidence of intended use and clinical risk.health ITMust align generative AI CDS development with the lifecycle-based approach.health policyShould follow how the competency-based approach, modelled on the way physicians are assessed, develops.Utah lets AI prescribe acne medication without direct physician oversightdoctorTwo physicians' approval is required for the first 100 prescriptions, so the supervisory role remains for now, but part of routine prescribing could shift to AI.health policyUtah's model is a template for the regulatory sandbox and AI prescribing, worth watching because of its relationship with the FDA.hospital managerIt is worth assessing how the reshaping of routine care affects outpatient volumes and liability questions.health ITIntegrating photo-based assessment and prescription issuing raises new vendor and data-handling requirements.California law says medicine must be practised by doctors and nurses, not AIdoctorThe law reaffirms that clinical decisions remain the physician's responsibility, even with AI support.hospital managerWhen deploying AI systems, hospitals must ensure that a qualified professional still makes the clinical decision.health policyThe contrast between the Californian and Utah approaches is a reference point for further state regulation.Utah approves AI system that prescribes medication without physician oversightdoctorThis is a precedent for handing the prescribing role to AI, so professional chambers will need to take a position on liability.health policyThe model is worth watching as a possible template or a cautionary example in regulating prescribing authority.pharmaceuticalsDirect AI prescribing could also change the access path for prescription-only products.AHA, AMA, ATA and radiology groups respond to FDA plan for generative AI medical devicesmedtech companyRisk-based classification, PCCPs and manufacturer accountability bear directly on the authorization plans for generative AI products.doctorThe professional bodies' proposals may also shape how clinician responsibility and oversight are divided.hospital managerDiverging FDA and state rules create procurement and compliance risk for generative AI tools.health policyThe conflict between state and federal levels creates an unclear regulatory environment.FDA and CMS scramble to avoid regulatory logjam as more states look to test AI in clinical settingshealth policyMisaligned state and federal rules could create uncertainty for AI pilots.hospital managerLicensing and payment conditions should be clarified before joining state AI experiments.Senate draft would let Medicare test paying for AI that performs clinical tasks without real-time physician reviewdoctorA payment model for AI without physician oversight raises questions of professional liability and division of tasks.health policyThe payment model could be the first major state-funded trial of clinical AI autonomy.medtech companyIt could open a reimbursement path for autonomous clinical AI tools.Stakeholders offer competing views on how FDA should regulate generative AI in medtechmedtech companyApproval strategies for generative AI products must be adjusted as the FDA's approach evolves.health policyIn shaping the regulatory framework, the challenge is balancing the differing interests of stakeholders.AI writes acne prescriptions without a doctor's visit in one US statedoctorAutonomous AI prescribing raises new questions of professional liability and patient safety.health policyThe rules on AI's authority to prescribe need to be clarified.
Machine analysis based on story titles and summaries, with a verification pass. The source of the facts is the article linked at each story.