New Uruguay/OKFN Partnership: Scaling Reliable AI for Open Data Across Latin America

Key Takeaways

• OKFN prototypes AI chat for Uruguay's open‑data catalog using MCP.
• Solution links natural‑language queries to CKAN datasets with traceability.
• Pilot aligns with Uruguay's 2024‑2030 AI and Data Strategies.
• Parallel Brazil pilot creates reusable blueprint for Latin America.
• Project stays tech‑agnostic, not tied to any specific LLM.

Pulse Analysis

The surge of open‑data portals across the globe has outpaced the tools citizens use to extract insight from them. Uruguay, long‑standing champion of digital government, has responded by embedding artificial intelligence into its National Open Data Catalogue. Partnering with the Open Knowledge Foundation, the country is testing a conversational interface that translates everyday questions into precise queries against CKAN, the open‑source platform that powers most public‑data repositories. This move not only modernizes access but also aligns with Uruguay’s 2024‑2030 AI Strategy, which emphasizes ethical, sovereign use of public information.

At the heart of the prototype is the Model Context Protocol (MCP), an open bridge that links large‑language models to trusted data sources while preserving provenance. Unlike black‑box AI answers, MCP‑enabled responses embed citations that point back to the exact dataset, enabling auditors and users to verify accuracy. The architecture is deliberately technology‑agnostic—compatible with any LLM and not locked to CKAN—so governments can adopt it without overhauling existing infrastructure. By foregrounding traceability, the project addresses a key barrier to public‑sector AI adoption: trust.

The Uruguay effort is part of a coordinated rollout with Brazil under the Digital Public Goods Alliance, creating a scalable blueprint for responsible AI in open‑data ecosystems throughout Latin America. If successful, the model could be replicated in the alliance’s growing membership, accelerating the region’s transition to data‑driven public services while safeguarding ethical standards. Funding from the Patrick J. McGovern Foundation underscores the philanthropic interest in democratizing AI for the public good. As more governments experiment with similar integrations, the MCP framework may become a de‑facto standard for transparent, citizen‑centric AI.