The Department of Health – Abu Dhabi establishes the world’s first regulatory foundations for researching and evaluating artificial intelligence systems in clinical tasks within a living lab model.
Launch of the smart clinical practice framework at the Abu Dhabi Future Health Summit
Evidence-based approach and safeguards for physician oversight and patient safety
Clinical risk variance and Abu Dhabi living lab trials
Medical liability governance and distinguishing between regulatory preparation and actual licensing
Frequently asked questions
Launch of the smart clinical practice framework at the Abu Dhabi Future Health Summit
On October 20, 2026, the Department of Health – Abu Dhabi unveiled a comprehensive regulatory framework dedicated to the research and potential future licensing of artificial intelligence systems to perform specific clinical activities that traditionally required a licensed healthcare practitioner. In a statement reported by the Emirates News Agency, the department described this framework as the first of its kind globally, while methodically noting that this comparative description originates from the regulatory body itself without the announcement including a comprehensive independent comparison with all international medical licensing approaches.
This new framework was reviewed during the Abu Dhabi Future Health Summit events by department chairman Mansoor Ibrahim Al Mansoori and undersecretary Dr. Noura Khamis Al Ghaithi. According to official coverage, the framework establishes an evidence-based approach to evaluate smart systems that may gradually perform broader categories of clinical work, with the department emphasizing the necessity of proving artificial intelligence proficiency in precisely defined tasks and providing stronger evidence as the proposed scope of the task expands.
Evidence-based approach and safeguards for physician oversight and patient safety
The regulatory framework outlines human physician oversight, clear accountability, and referral to a specialist when appropriate as indispensable core safeguards. These stringent requirements distinguish between regulated research in clinical artificial intelligence and granting unrestricted software authorization to practice medicine completely independently.
As of the announcement date of October 20, 2026, this framework represents a foundational base for future policies and supporting standards rather than a public list of artificial intelligence systems that have already obtained licenses to operate as independent clinical practitioners; its primary objective is to govern the research and development path prior to any operational adoption.
Clinical risk variance and Abu Dhabi living lab trials
Clinical activities vary in their levels of medical consequences and risks; a software tool that summarizes a medical appointment carries completely different risks from a tool that selects a treatment plan or provides advice regarding emergency and critical symptoms. Therefore, evaluating any system requires defining the precise task, the target patient category, acceptable error rates, the types of data used, and the conditions that mandate immediate human intervention, as general model performance in medical exam questions is not sufficient evidence of safety in every clinical role.
The Department of Health – Abu Dhabi plans to conduct this research within the emirate’s “living lab” model, relying on its healthcare infrastructure and organized research pathways that allow precise monitoring under controlled conditions before wider operational decisions are made. However, the statement did not specify the full test population, approved companies, commercial contracts, or the specific date for granting any system expanded authorities.
Medical liability governance and distinguishing between regulatory preparation and actual licensing
This proposal raises fundamental questions regarding legal and medical liability; when artificial intelligence advice is incorrect, the healthcare institution needs clear records documenting what the program did, the specialist physician’s review, and the circumstances that necessitated case escalation. The success of the framework—which requires stronger evidence for activities with greater consequences—relies on detailed implementation standards and actual enforcement.
Abu Dhabi’s announcement gains particular importance because most current health artificial intelligence tools provide decision support without claiming the authority to execute a licensed clinical task; thus, the framework examines conditions under which these boundaries might be reviewed while maintaining patient safety conditions, without reducing it to misleading headlines claiming the current replacement of human physicians, pending published technical standards and approved study protocols in the future.
Frequently asked questions
Question: What is the new framework unveiled by the Department of Health – Abu Dhabi?
Answer: During the Abu Dhabi Future Health Summit, the department unveiled a regulatory framework for research and potential future licensing of artificial intelligence systems to perform specific clinical tasks that traditionally require a licensed healthcare practitioner.
Question: Does this framework mean immediate permission for artificial intelligence programs to practice medicine without a doctor?
Answer: No, the framework represents a foundational base for future research and policies, requiring physician oversight, accountability, and medical referral as essential safeguards, and does not grant absolute authorization for software to practice medicine independently.
Question: How will the Department of Health – Abu Dhabi test clinical artificial intelligence systems?
Answer: The department plans to conduct research and evaluations within the Abu Dhabi living lab model using its healthcare infrastructure to monitor systems under controlled conditions prior to any operational expansion.
Question: What are the required criteria to evaluate the safety of any clinical artificial intelligence system?
Answer: Evaluation requires precisely defining the clinical task, patient category, acceptable error rates, data type, human intervention conditions, and providing stronger scientific evidence as the task becomes riskier.
