Latin American open data communities, governments, and civil society are gathering this week in Guatemala City for Abrelatam/ConDatos 2026, the region’s most important open data conference at a pivotal moment for democratic governance. Transparency, responsible access to information, and civic participation are key pillars to guaranteeing and strengthening democracy in such uncertain times. Since 2013, Abrelatam/ConDatos has served as a critical space for articulating shared agendas across sectors.

This year, Open Knowledge Foundation (OKFN) is proud to contribute once again to that tradition.

We are bringing some of the results of our AI Learning Labs initiative to the regional stage, developed in collaboration with key partners in Latin America.

The AI Learning Labs are designed to turn real-world experiments into open, replicable methods for responsible AI, so that any public-interest organisation can build AI literacy without reinventing the wheel. Rather than producing abstract training materials, we develop replicable methods that help organisations build AI skills, use AI responsibly, and develop their own AI projects.

At Abrelatam/ConDatos 2026, we will share learnings and strategies from two regional pilot projects that sit at the heart of the initiative. 

The first was developed with the Governments of Brazil and Uruguay through CGU and AGESIC. Together, we have prototyped an AI integration with open data portals that enables users to ask natural-language questions and receive responses traceable directly to the underlying datasets. This work builds a bridge between large language models (LLMs) and public open data catalogues in a way that ensures all information is reliable and trustworthy.

The second focuses on digitising at-risk physical archives. In partnership with AVANCSO (Association for the Advancement of Social Sciences in Guatemala), we have been exploring how to preserve a unique documentation centre that holds historical records of social and indigenous movements for human rights through decades of war in Central America. Anthropologist Alejandro Flores-Aguilar has been developing a plan for how to preserve this knowledge for future generations, including by using AI, while navigating questions of open source platforms, sustainability, governance, security risks, and technological sovereignty.

Both sessions are designed as open conversations. We invite everyone attending Abrelatam/ConDatos to come along, ask questions, and explore opportunities for collaboration.

New publications edited by OKFN will be launched to accompany each of the above sessions. These field guides are designed to help participants replicate our AI experiments in their own contexts and with their own resources.

Attendees will be able to collect printed copies. Links to download the digital versions are provided below.

‘Connecting AI to Government Data’ (2026)

How can we use artificial intelligence to make public data more accessible without compromising accuracy or trust? Key content in the guide includes a pilot ‘recipe’, six steps to creating a data provenance framework for AI agents, practical recommendations for people and machines, and a reflective section acknowledging the limits of experimentation.

‘A Library That Supports Research Into a Collection’ (2026) (Spanish version only, English coming soon)

These pages are aimed at those working in documentation centres who are wondering what is needed to prepare a collection for AI applications. The text offers a starting point for organising the initial decisions for a pilot project. The guiding question is: what would a centre need to do to ensure that an application helps to research its materials and allows users to trace back to the documents and pages underpinning its results?

Open Knowledge Foundation’s representative on the ground in Guatemala is Andrés Vázquez, Senior Software Engineer. Andrés has been a driving force behind the technical development of the AI Learning Labs, working directly with our partners in Latin America to build open, traceable AI solutions for public data.

If you are attending Abrelatam/ConDatos, we warmly invite you to connect with Andrés. He is an expert and long-time contributor to CKAN’s core code, and is responsible for developing the sandbox connecting our MCP experiment to the International Aid Transparency Data (IATI) datasets, open for anyone interested in testing the tool.

Whether you want to discuss the MCP architecture, explore how our work could support your organisation’s work, or simply share ideas over a coffee – he is looking forward to connecting the Latin American open data community in person. New projects, new partnerships, and new questions are all welcome. Come and say hello!

• On LinkedIn: https://www.linkedin.com/company/open-knowledge-foundation/
• On Bluesky: https://bsky.app/profile/okfn.bsky.social
• On Mastodon: https://fosstodon.org/@okfn
• On X: https://x.com/okfn

Media contacts: lucas.pretti@okfn.org in English, Spanish, and Portuguese. For any other aspects relating to OKFN’s participation at Abrelatam/Condatos, contact info@okfn.org.