The Future of AI Enabled Health 2025

Technical choices are often the remit of technical experts, and leaders tend to shy away from engaging in technical matters. However, this is a dangerous pattern, given how some technical decisions limit or constrain the ability to deliver on a vision. Historically, due to legacy systems, persistent budget constraints and an undersized workforce necessitating strategies that optimize productivity and effectiveness, health systems have evolved through a series of independent tenders and cost-oriented procurement processes with no real overarching strategic perspective. This has led to disjointed infrastructure, even within an organization; this is often the case in hospitals. To move beyond this ad hoc development, a clear, long-term architectural vision is essential. This vision should incorporate health needs into broader digital public infrastructure (DPI) strategies to ensure that stakeholders (1) do not attempt to “reinvent the wheel” and (2) secure equitable access to AI-driven health solutions. Indeed, the persistent access gap to digital health solutions due to inadequate DPI further exacerbates the challenge. For example, focusing on comprehensive datasets that include social determinants of health (SDOH) can help address biases and privacy issues in AI systems. Promoting local production initiatives and ensuring that datasets are reflective of local contexts can bridge the digital divide and support the implementation of AI-driven health solutions in underserved regions. By managing legacy systems and integrating them with new AI technologies, health systems can support AI adoption without discarding valuable existing infrastructure. This comprehensive approach ensures that health needs are met within a broader digital strategy, promoting equity and efficiency in AI deployment. Only through a combination of intentional architectural planning and enhanced digital literacy can health systems effectively use AI to deliver comprehensive, equitable care. Persistent inequitable access to digital health solutions due to inadequate DPI Equity must remain at the forefront of AI initiatives in health, and efforts should focus on ensuring that AI tools increase access for those most in need, bridging the gap between populations with and those without quality health services.Table 2 summarizes the main points and strategies related to equity in AI initiatives in health, categorized into data ownership/access, public goods/infrastructure. Each category addresses different aspects and challenges in ensuring equitable access and the effective use of AI technologies in health. Successful examples of addressing data ownership and DPI gaps include the United Kingdom’s NHS and Israel’s health sector. In the United Kingdom, the NHS centralizes public and private health data to enhance transparency and benchmarking, while Israel’s centralized data management system played a crucial role in its successful COVID-19 vaccination rollout. By early January 2021, Israel had vaccinated more than 14% of its citizens, outperforming many larger and wealthier nations. These examples show that effective data management and the integration of health needs into broader digital infrastructure are essential for achieving large-scale health success. Such integration not only enhances transparency and efficiency but also ensures that health systems are better prepared to respond swiftly and effectively to public health challenges such as pandemics. Limited digital literacy is hindering leaders and decision-makers The importance of digital literacy among key stakeholders cannot be overstated. AI is a relatively new field, and many decision-makers have not had the opportunity to become well versed in it. However, a lack of fundamental knowledge about AI slows progress. This limited understanding can spark societal fears, lead to regulatory missteps and result in a hesitation to embark on the AI journey. To bridge this gap, there is a pressing need for comprehensive upskilling. Leaders must grasp not only the strategic purposes of AI but also the foundational digital principles that underpin it. Defining the minimum required knowledge for informed decision-making regarding AI is crucial. Ensuring that all stakeholders are on the same page will facilitate smoother discussions and more effective implementation strategies. Equipping decision-makers and AI users with the right approach to interrogate AI solutions – tracking key performance indicators and clinical outcomes, for example – will help them understand the value of AI and generate trust.3.2 Misalignment of technical choices with strategic visions The Future of AI-Enabled Health: Leading the Way 18