Center for DPI unveils framework for AI-ready nations

The Center for Digital Public Infrastructure has published a paper setting out what it considers a vision for “Building AI-Ready Nations through Digital Public Infrastructure.”

It presents the DPI-AI Framework, which is a guide to thinking practically about how AI can be integrated through DPI.  The paper focuses on digital identity, data exchange and payments, which are core parts of DPI. The framework lays out three complementary elements: AI Blocks, DPI Workflows and Public Agents.

The paper is aimed at government, development partners, donors and practitioners shaping local, national and international digital agendas. It outlines ways to align emerging AI capabilities with DPI towards development outcomes.

AI blocks are modular units of AI capability, which can perform sector-specific functions such as identity verification or credential issuance, as well as broader tasks like classification and translation. DPI Workflows are the orchestration layer that slots AI Blocks into existing DPI systems. Public Agents are AI-driven interfaces enabling interaction with citizens and public servants.

“Together, these elements describe how intelligence can be embedded into public service delivery while preserving interoperability, accountability, inclusion, and sovereignty,” the paper says. The paper believes AI fits within DPI as another modular building block and that it is not necessary to create a new model, emphasizing DPI principles of minimalism and reusability.

“Just as APIs enabled interoperability in earlier generations of digital infrastructure, emerging open standards such as Model Context Protocols (MCPs) now make it possible for AI systems to interconnect through shared context layers,” the paper argues.

The Building AI-Ready Nations through Digital Public Infrastructure paper also contains an 11-point list of challenges. These run through safeguards, procurement and governance. For example, the verification and certification of AI Blocks presents a dilemma.

“AI Blocks can change behaviour over time, making one-time approval insufficient,” it explains. “Clear responsibility is needed for who validates AI Blocks, against what criteria, and with what ongoing oversight.” It suggests continuous certification as one approach. Another challenge is defining public versus private AI Blocks.

However, there are opportunities in accelerating public innovation, in faster and smarter responses and services, and in distributing intelligence rather than it being centralized. The full vision paper from the Center for Digital Public Infrastructure can be read here.

Article Topics

data exchange platform  |  digital identity  |  digital payments  |  digital public infrastructure  |  DPI Safeguards Framework  |  DPI-AI Framework