Digital Public Infrastructure (DPI) is increasingly recognised as a core pillar of digital transformation at national and organisational levels, providing platforms for the development of additional services that benefit the public. As DPI adoption accelerates, there is a growing need to ensure that rapid development and deployment are carried out responsibly, with appropriate governance, safeguards, and institutional oversight in place. In the advent of AI, it is clear that it can significantly enhance DPI by enabling among other things multilingual interfaces, speech recognition technologies, personalisation, fraud detection, and predictive analytics capabilities that can dramatically improve accessibility and user experience, particularly in diverse and low-resource contexts. However, these AI-enabled capabilities also introduce new complexities that must be carefully examined to build trust and accountability around DPI systems. Prototyping and experimentation tools such as regulatory, policy, and technical sandboxes can provide a controlled setting to test and understand the implications of new technologies and data practices. If designed effectively, sandboxes can foster trust and accountability by involving diverse perspectives in the development of standards for privacy, security, ethical use, and risk mitigation.
The AI-DPI experimentation and the role of Sandboxes session will engage participants through a combination of presentations and a panel discussion, with a focus on how experimentation can support safe, accountable, and trustworthy AI adoption in DPI systems. The conversation will explore among other things the challenges that affect DPI systems and the complexities introduced by AI through experience of practitioners from government and development organisations.
Topics will include:
The session builds on the AI-DPI Sandboxing Official Pre-Summit Event (December 2025), organised by Datasphere Initiative, Kalpa Impact, and UNDP, where experts identified key challenges for AI-DPI experimentation.
In the end, the session aims to advance safe, inclusive DPI innovation, generating global governance insights, design principles, and sandboxing strategies for the responsible design and deployment of trustworthy AI-powered DPI.
Important: Session registration is separate from summit registration. Please ensure you have registered for the full summit to attend.
