“Artificial intelligence (AI) holds great promise for improving healthcare and medicine worldwide, but only if ethics and human rights are placed at the very core of its design, deployment and use.” (World Health Organization – Ethics and Governance of Artificial Intelligence for Health – 2021)
By Azir Aliu
The healthcare system in North Macedonia is not at the beginning of integrating digitalization. The process has already started, has been institutionally established and continues to develop systematically, and most importantly, it is visible. “My Appointment” has, for more than a decade, connected electronic health records, referrals and prescriptions. The “My Health” application enables citizens to access their examinations, referrals, prescriptions and vaccinations. In parallel, legal amendments have been prepared that foresee the full electronic management of medical documentation, while telemedicine is expected to gain equal institutional value to physical examinations. North Macedonia is already intensively building a digital foundation, and therefore a more important question now stands before us: will we enter the era of AI prepared, or will we allow it to catch us without a prepared healthcare, legal and ethical response?
First, it is necessary to emphasize a fundamental distinction that is often lost in public debates. Digitalization collects, connects and organizes data, while AI analyzes, predicts, recommends and begins to influence reasoning and decision-making. Digitalization is the infrastructural base, while AI is the upgrade—the new phase that absorbs the power of decision-making. Therefore, North Macedonia does not need only a general digital transformation of healthcare, but the development of a specific approach to AI. What is required is a certain level of state maturity, expressed through a nuanced systemic framework that brings together medicine, law, ethics, cybersecurity and the public interest.
International institutions are already warning that this is precisely where the most sensitive gap emerges. In 2025, the WHO found that only 8% of European countries have a dedicated national strategy for AI in healthcare, while 86% cite legal uncertainty as the main barrier, and only 8% have established accountability standards if the system causes harm. The European Commission, meanwhile, notes that software for medical purposes falls into the category of high-risk systems and must therefore be subject to strict requirements regarding data quality, human oversight, risk mitigation and clear information for users. The essence is that technology advances rapidly, but trust in healthcare must not be built rapidly—it must be developed carefully and correctly.
This issue also carries a deeper philosophical and ethical layer—fear. In this case, it does not stem only from AI as something new and unfamiliar to our professional experience. It also arises from the possibility that the human being may fade into an unclear or blurred factor within a system that is supposed to be crystal clear—the patient may not know who makes the decision, while the doctor may feel that an invisible mechanism is inserted between their knowledge and the fate of the patient. This is precisely why UNESCO emphasizes that the protection of human rights and dignity is the foundation of AI ethics, and that human oversight must not be reduced to a procedural formality, but must be a substantive guarantee. Technological enthusiasm alone is not enough to reduce fear; what is needed is a moral structure within the system. People will trust AI in healthcare only if they are certain that it does not remove the human being from the center.
For this reason, it is important to state something simple: in medicine, there must be no such thing as an “anonymous decision.” If an algorithm assists in triage, screening, interpretation of imaging, or risk assessment, this must be verifiable and explainable. If the system makes an error, responsibility must not, under any “objective” circumstance, disappear into the grey zone of algorithmic calculations. The risk of biased decisions may stem from the bias of the data on which those decisions are based. In this context of potential and real risks, a directed and careful approach to integrating AI becomes essential—one that continuously corrects processes in healthcare in order to keep them within ethical boundaries.
The world already offers concrete examples that AI can function as part of a healthcare system, but only when placed within a clinical and institutional framework. In England, the National Health Service (NHS) reports that a tool for rapid interpretation of brain scans in stroke cases has been deployed across more than 70 hospitals. Around 15,000 patients have directly benefited, and thanks to faster detection of clots, life-saving intervention is delivered more than an hour earlier. This is not science fiction—it is a concrete example where AI is embedded in a specific healthcare segment—acute stroke care—to help doctors make faster decisions where every minute matters for the outcome of treatment.
North Macedonia, for its part, already has processes indicating that this issue is being addressed in a timely manner. The Ministry of Digital Transformation last year presented the “Vezilka” project—a National Center for Artificial Intelligence—where healthcare is explicitly listed as one of the priority areas, particularly through the analysis of health data and the improvement of medical services. At the same time, the Ministry of Health is working on finalizing a National Digital Health Strategy, focusing on interoperability, HL7 FHIR standards, audit trails, role-based access and real-time analytics, while a telemedicine pilot project has already been successfully implemented on the Skopje–Debar–Delchevo route. This means that the country is already opening the door to the next phase. But every open door requires a guardian—and in healthcare, that guardian must be national policy.
For this reason, I believe that our country needs to develop its own credible approach to AI in healthcare—an approach in which the final clinical decision is made by a human, not by an algorithm. Within such a framework, every AI tool must be clinically validated, ethically supervised and legally regulated before becoming part of practice. Sensitive health data must not be turned into raw material without rules and strong protection. Doctors must be trained to use technology, but also to understand where its utility ends and where human judgment and decision-making begin. Most importantly, the patient must be assured that the state has not sacrificed their safety in the name of the speed of innovation.
New digital tools and AI are important, but they are not sufficient to fully determine the future of healthcare. Human wisdom is needed to place them within rules. Digitalization is already an intensive process in North Macedonia, while the next major step will be the creation of a distinct, responsible and dignified approach to AI in healthcare.
(The author is Minister of Health of North Macedonia)