Artificial intelligence could play an important role in helping Asia-Pacific healthcare systems manage growing workforce pressures, noted by Ilya Burkov, Global Head of Healthcare & Lifesciences at Nebius, during an interview.
Burkov noted that the region is facing a major demographic shift, with ageing populations increasing the demand for care. Japan has around 30 per cent of its population aged over 65, Singapore has officially become a super-aged society, and South Korea is facing a general practitioner shortage that has reached crisis point.
At the same time, non-communicable diseases such as heart disease, diabetes and chronic obstructive pulmonary disease are becoming a growing burden across Asia and the Pacific. According to Burkov, these conditions are now the leading cause of death across the region, accounting for more than half of all deaths in every country.
These pressures are contributing to potentially avoidable hospitalisations, in which patients end up in hospital for conditions that could have been prevented or managed earlier with timely routine care. Burkov said this shows that hiring more specialists alone may not be enough at the pace the region needs. Instead, the focus should be on extending the capabilities of the existing healthcare workforce.
One area where AI could support this shift is in enabling non-specialists or minimally trained staff to perform selected tasks that have traditionally required specialist involvement. Burkov pointed to pilots involving AI-guided echocardiograms, in which staff without clinical training can operate the system, while AI automatically captures measurements and clinicians review the results remotely.
This type of model does not remove the need for clinicians, but it could help reduce specialist bottlenecks and improve access to diagnostic services. Burkov described AI as a capable junior assistant that handles defined and repeatable tasks, while flagging uncertain cases for human review.
However, he stressed that supervision remains essential. AI-supported care models work best when there is clear oversight, a defined scope and a clinician accountable for the final decision. The technology may remove the bottleneck, but it does not remove clinical accountability.
For healthcare systems to scale these models safely, governance needs to be built before deployment. Burkov said different levels of AI risk should be treated differently, with low-risk decision-support tools not regulated in the same way as high-risk independent systems.
Clinical AI also needs to be designed with auditability from the start. This includes structured outputs, traceable sources, model version control and documented evidence that guideline-directed recommendations can be reproduced and reviewed. These safeguards are important in building trust among clinical teams.
Burkov also highlighted that successful AI implementation should not be treated as only an IT project. Hospitals need to involve clinicians, nurses, AI experts, ethics leads and compliance professionals from the beginning. Even a technically accurate tool may fail if it adds friction to a fast-paced clinical environment or creates outputs that healthcare workers cannot interpret quickly.
Reimbursement and incentive structures are another major barrier. Burkov noted that many healthcare systems still pay for treating sick patients, rather than keeping them well. This creates challenges when an AI tool prevents an avoidable hospital admission, but the financial savings happen in one department while the cost of the tool sits in another.
To support wider adoption, healthcare systems may need more holistic economic models that reward improved outcomes across the full care pathway. Value-based models could create stronger conditions for prevention-first technologies to attract investment and scale more effectively.
Looking ahead, Burkov expects diagnostic workflows to see some of the earliest impact from AI, especially in areas such as cardiac diagnostics, radiology and pathology. These are high-volume, pattern-recognition-heavy fields where demand is high and specialist availability can be limited.
Chronic disease management could also see a wider population-level impact. AI could help with risk stratification, identifying patients who may deteriorate earlier, and personalising follow-up frequency instead of relying on one-size-fits-all care models. Administrative tasks such as synthesising patient records, flagging diagnostic gaps and supporting guideline compliance could also free up clinician time for higher-value patient interactions.
For ageing societies such as Japan, Singapore and South Korea, AI could support a shift away from care models concentrated mainly in specialist facilities. By extending diagnostic and monitoring capacity into community settings, pharmacies, homes and primary care environments, AI could help patients receive care earlier and closer to where they live.
Burkov said the combination of strong digital infrastructure and urgent workforce pressures creates favourable conditions for AI adoption in these markets, provided governance frameworks and economic models keep pace. As APAC healthcare systems continue to face ageing populations and rising chronic disease burdens, AI could become a constant support layer across the patient journey, helping clinicians prevent disease earlier, customise long-term care and guide patients to the right specialists faster.
