Somewhere in rural Kenya, a community health worker uses a mobile application to screen a child for malaria. In a busy outpatient clinic in Shenzhen, a patient describes symptoms to an AI-powered system before ever seeing a physician. In a rehabilitation ward in Baltimore, a pharmacist receives an automated alert about a potentially dangerous drug interaction in an elderly patient’s prescription. These are not hypothetical futures. They are part of a present reality: artificial intelligence is steadily transforming healthcare across the world.
For diplomats, policymakers, healthcare leaders, and international organisations, the question is no longer whether AI belongs in medicine. It does. The more meaningful question is how to ensure that this transformation reaches everyone, not only those who already benefit from the most advanced healthcare systems.
AI-enabled healthcare is making a practical difference in multiple ways. Pre-consultation tools, for instance, are helping address one of medicine’s most persistent frustrations: the waiting room. When patients provide symptom information and relevant medical history before a consultation, clinicians can arrive better prepared. The result is not an effort to replace clinical judgment, but to improve the quality and efficiency of clinical encounters. This approach is proving effective across very different healthcare contexts. High-income systems use AI-assisted triage and workflow tools to improve efficiency and clinical compliance, while lower-resource settings adopt similar approaches to maximize scarce healthcare capacity. In both cases, the objective remains consistent: enabling patients to communicate effectively and enabling healthcare professionals to focus on the decisions that matter most.
AI is also contributing to addressing specialist shortages. In parts of Africa, Southeast Asia, and Latin America, limited access to paediatric specialists remains a significant barrier. Diagnostic support tools are helping narrow this gap in practical ways, including through solutions that can operate on basic smartphones and, in many settings, with limited or intermittent connectivity. Rather than attempting to replicate specialist expertise everywhere, these systems support frontline health workers in making safer determinations about whether a child needs urgent referral or can be managed locally. Where specialist care may be hours away, these decisions can directly shape outcomes.
In parallel, AI is increasingly relevant to safer prescribing and the management of chronic conditions. Polypharmacy—the simultaneous use of multiple medications—creates substantial risk for harmful drug interactions, especially among older patients. AI systems capable of analysing extensive compatibility rules and patient-specific factors can provide real-time alerts. Importantly, these alerts do not eliminate the need for clinicians; they strengthen their decision-making by helping them account for complex interactions that would be difficult to track reliably in a busy setting.
Perhaps most importantly, AI is helping bridge the “last mile” to communities historically left behind. Offline diagnostic and screening tools, when appropriately designed and deployed, can support disease detection, prescription review, and public health reporting in locations where specialist expertise is scarce. These solutions may not always appear revolutionary, but they matter precisely because they extend healthcare capacity to where access has traditionally been weakest.
And yet, even with these demonstrated benefits, AI-enabled healthcare does not scale as rapidly as it should. The technology exists. The need is evident. Still, multiple barriers remain. One of the most prominent challenges is regulatory fragmentation. There is no universally recognised framework for evaluating and approving AI-enabled healthcare products. A diagnostic tool that has been rigorously validated in one jurisdiction may encounter entirely different approval processes elsewhere, with new requirements, timelines, and costs. The consequence is predictable: deployment slows, innovation becomes harder to finance, and the regions that could benefit most are often the last to receive proven tools.
Even when products obtain regulatory approval, market access can remain uncertain. Reimbursement and procurement pathways are not always clearly established, and health insurance systems may struggle to integrate AI-enabled services into routine care. Without sustainable financing mechanisms, health ministries find it difficult to invest with confidence, while innovators struggle to build business models that can extend beyond high-resource markets.
Data silos form another structural obstacle. AI systems depend on high-quality data, yet healthcare data often remains fragmented across institutions. Inconsistent governance and regulatory approaches across countries can further complicate responsible data sharing. Alongside technical challenges, there is also a trust deficit. Patients and healthcare institutions increasingly demand clarity about how data is collected, stored, protected, and used. Without trust-building and reliable governance of data ecosystems, even well-performing models may struggle to be adopted responsibly.
Implementation itself is frequently the most underestimated barrier. A technically successful AI solution can still underperform if healthcare workers are not trained, if infrastructure is inadequate, or if local communities are not meaningfully engaged. Deployment is ultimately a socio-technical challenge, one shaped by human capability, institutional readiness, and everyday clinical realities.
These themes were examined in depth during the Geneva School of Diplomacy’s session at the WSIS Forum 2026, “Breaking Barriers: Pathways to Accelerate AI-Enabled Healthcare.” The discussion brought together experts from healthcare, technology, policy, regulation, and international governance communities to consider what it will take to scale AI-enabled healthcare responsibly and equitably.
A strong consensus emerged: scaling AI-enabled healthcare is not merely a technological endeavour. It is simultaneously a trust and capacity challenge, a governance and interoperability challenge, and a patient-centred implementation challenge. In other words, the path forward requires collective responsibility across the healthcare ecosystem.
One of the most constructive messages was that AI should not primarily be treated as a threat. Healthcare rightly demands caution, but fear alone can slow adoption and reinforce misunderstandings. Participants argued instead that AI should be viewed as an investment—an investment that must be matched by sustained investment in education, implementation readiness, and public confidence. Healthcare professionals, regulators, policymakers, and health system managers need practical, role-specific education that equips them to understand what AI can do, what it cannot do, and how to interpret outputs responsibly. Human judgment remains essential, but it cannot be exercised well without adequate capability and clear operational guidance.
Trust-building, participants noted, is not created through marketing narratives. Public confidence is earned through transparent governance, rigorous evaluation, accountability mechanisms, and accessible pathways for addressing harms when they occur. This trust is inseparable from implementation readiness. Even strong technologies can fail in practice if they do not integrate with clinical workflows, if training programmes are missing, or if local support capacity is insufficient. The central takeaway was therefore straightforward: investing in AI requires investing in people and in trust.
Governance was a second recurring focus. Regulatory fragmentation was repeatedly described as one of the greatest obstacles to cross-border scale. Participants emphasised the importance of interoperable governance frameworks—approaches capable of harmonising regulatory expectations across jurisdictions, supporting interoperability of standards and technical requirements, and enabling cross-border recognition of validated AI-enabled healthcare tools. When governance frameworks are interoperable, mutual recognition becomes possible, duplication reduces, and proven solutions can be adapted more efficiently and safely across settings.
Inclusion, however, was not treated as a rhetorical principle. The session underlined that co-design and co-ownership must be structural. Developing countries should not be positioned merely as participants in global discussions, while policy and standards are largely shaped elsewhere. Instead, they should actively co-design and co-own governance frameworks, the processes through which standards are set, and the mechanisms that enable implementation. Participants noted that previous globalisation waves often resulted in the Global South serving mainly as a recipient of externally designed policies and technologies. AI-enabled healthcare offers an opportunity to do this differently—toward shared governance and shared responsibility, rather than asymmetrical rule-setting.
The discussion also clarified that the hardest problems often surface after a system has proven technically effective. AI-enabled healthcare must be implemented as part of a broader socio-technical system that includes clinical workflows, professional standards, financing mechanisms, health policy, and patient experience. For that reason, participants advocated for principles-based multistakeholder collaboration, bringing together health ministries and regulators, clinicians and professional associations, standards and conformity assessment bodies, civil society and patient advocates, innovators and digital health providers, and insurers and payers where relevant. The point was not simply that many stakeholders should be “consulted,” but that the system itself is too complex for any single actor to manage alone.
This leads to a crucial governance and design lesson: patients and clinicians must be part of the process. They should not be treated as passive recipients of AI-enabled services. Their perspectives should inform design, deployment, and governance because real clinical needs are best articulated by those who live with the consequences of failure and the value of improvement. Meaningful end-user involvement supports appropriate communication and consent practices, helps identify safety concerns early, and prevents AI deployment from adding burdens to communities that are already vulnerable.
Finally, participants highlighted collective responsibility beyond clinical performance. As healthcare platforms become more connected, they expand potential attack surfaces. Scaling AI-enabled healthcare therefore requires security-by-design approaches and coordinated incident response capabilities, supported by clear cybersecurity governance expectations across the ecosystem. Evidence generation must also be continuous rather than one-off. While many AI tools perform well in controlled settings, participants stressed the need for evidence across diverse populations, across different healthcare system designs, and in real-world operating conditions. Continuous monitoring, post-deployment evaluation, and ongoing learning are needed to ensure that AI remains safe, effective, and equitable over time. In short, safety must include both cybersecurity resilience and continuous evidence generation.
Taken together, the outcomes from the GSD WSIS Forum 2026 session point toward a coherent agenda for international cooperation. It requires investing in education, trust-building, and implementation readiness. It calls for interoperable governance and harmonised standards that enable cross-border scale. It demands co-design and co-ownership by developing countries so that governance is genuinely shared. It asks for multistakeholder, patient-centred implementation models that reflect healthcare realities rather than technology assumptions. And it requires collective responsibility for cybersecurity and evidence gaps so that safety and effectiveness can be sustained.
No single government, company, institution, or international organisation can scale AI-enabled healthcare on its own. Achieving meaningful, equitable impact depends on shared rules, shared accountability, and a shared commitment to improving health outcomes worldwide. This is precisely the type of challenge for which Geneva’s multilateral ecosystem is well-suited. Effective diplomacy has always involved bringing together diverse stakeholders to address problems that transcend borders. The future of AI-enabled healthcare is one such challenge, and it demands collective action.
Acknowledgements
The Geneva School of Diplomacy extends its sincere appreciation to all panellists, participants, partners, and contributors who supported the session “Breaking Barriers: Pathways to Accelerate AI-Enabled Healthcare” at WSIS Forum 2026. Their insights on education and public confidence, interoperable governance, patient-centred implementation, cybersecurity, evidence generation, and collective responsibility helped shape a constructive and forward-looking agenda for scaling AI-enabled healthcare globally. This report is not reviewed and edited by all panellists; thus, all errors are of authors.
