When end users can build: AI-assisted application development and the new accountability gap in digital health

Generative artificial intelligence (AI) is dramatically changing the division of labor in digital health innovation. Until recently, a frontline health professional, educator or researcher who identified a workflow problem typically needed a software developer to translate that problem into a functioning application. Drawing on my recent experience building and deploying two web applications for health professional education, each in less than five hours and without formal computer science or information technology training, this Perspective argues that the early bottleneck in digital health tool development is shifting. AI did not replace domain judgment: the clinical education needs, workflow priorities, and decisions about what counted as useful remained mine. What AI supplied was technical translation, code generation, troubleshooting and step-by-step guidance for using tools such as Visual Studio Code, PowerShell, version control and web deployment platforms. This experience suggests that end users may increasingly be able to move from problem identification to first functioning application without a software developer at the initial build stage. The implications are substantial but not simple. Technical collaboration remains essential, but its center of gravity may move toward cybersecurity, privacy, architecture, institutional integration, maintenance and long-term stewardship. The unresolved issue is not whether end users can now build, but who is responsible for determining whether AI-generated code is safe enough to deploy in institutional environments. Digital health practitioners and researchers therefore need governance models that recognize a new reality: end-user-developed applications are becoming technically feasible, but accountable review and ownership remain indispensable.