Future outlook: Four visions for AI-enabled health2 Expert visions of AI’s future in health balance enthusiasm with a measured scepticism. To better grasp the transformative potential of AI in health, it is essential to explore the varied perspectives of experts in the field. Their insights offer a glimpse into the future, highlighting opportunities and challenges. The experts interviewed (Figure 5) provided invaluable insights from leading organizations in AI and health. They were asked for their input on four AI-driven hypothetical visions that are summarized in Figure 6. These are not mutually exclusive or definitive; instead, they illustrate what could be possible with AI. While some health professionals were enthusiastic, others were cautiously sceptical, highlighting the fact that despite the optimism, significant technological and structural constraints remain. Current large language models (LLMs), for example, still face issues with hallucinations,9 or non-existent patterns or objects, indicating that this generation of AI tools still needs to mature. Additionally, there are substantial technological gaps, especially in LMICs. Understanding and addressing these nuanced challenges is essential for unlocking AI’s full potential in health. The fact that these visions were not proposed a decade ago is not rooted in technology but in systemic structural constraints. Describing the visions helps identify the constraints and challenges. The visions highlight the transformative potential of a complex landscape of traditional AI (e.g. machine learning) and genAI. Incorporating both static and dynamic approaches poses unique opportunities and challenges, including the need for tailored regulatory and strategic approaches. Expert group composition FIGURE 5 Source: Digital Healthcare Transformation Initiative dialoguesI nternational organization G over nment Think tank H ealthcar e/tech company F oundation N on-gover nmental organization (NGO) R esear ch institution S ocial enterprise15 5 5 28 4 9 7 2 0 2 0 4 0 6 0 8 0N umber of participants Total7 5 The experts considered these four AI-driven visions: –Transformation in well-being: This vision highlights a widespread availability and use of sensors to generate extensive data, enabling predictive care, lifestyle management and personalized wellness programmes. Implications include a shift in the economic model of health from treatment to prevention and workforce realignment to focus more on preventive care. The Future of AI-Enabled Health: Leading the Way 13

The Future of AI Enabled Health 2025