Project summary

Given that currently available AI tools are designed and conceived by labs, conglomerates and companies that are mandated to extract data as a matter of principle, the question of indigenous knowledge and their ancestral knowledge remains particularly at risk of complete loss, deterioration and distortion in every single aspect regarding their legacy.

This is exaclty what this initiative is meant to tackle. Still very much in its pre-seed stage, the core of this project, promoted and initiated by Hemispher-IA, seeks to develop an offline-first, mesh-networking platform for the Ngäbe-Buglé people across the Panama-Costa Rica border. It intentionally builds on a previous experience created to build precise localisation of the plight of communities affected by the long-lasting effects of internal displacement and violence in Colombia (Recordari, Universidad del Rosario, 2017).

This model and architecture are to be adapted and transformed into a newly introduced push to record and govern indigenous communities’ own cultural data.

With strict observance of community-centered and grassroots-protected data sovereignty, this is likely to solve questions relating to connectivity or administration by foreign data centres.

In the course of the coming 18 months, the proposed pilot will run in 10 to 12 communities across the common border shared by northwestern Panama and Southern Costa Rica.

Such a model will require ownership of local communities, with and estimated six (6) young so-called “Digital Weavers” from the Ngäbe-Buglé cross-border community. The goal is not just local preservation alone, but rather a technified platform of open-source, reusable tools to prevent digital extractivism from taking place within the outlined indigenous territories.

What are this project's goals? How will you achieve them?

Three main objectives are considered the absolute minimun at the project’s starting point.

First, observe and implement the CARE Principles, as described below:

(a) Collective benefit: Data ecosystems must enable Indigenous Peoples to derive value, support innovation, and foster community-led development.

(b) Authority to Control: Indigenous rights are fully recognized, empowering their communities and projected beneficiaries to control how their cultural knowledge, languages, and lands are represented and accessed.

(c) Responsibility: Those working with indigenous data have a duty to build positive relationships, support community capability, and ensure data usage honors indigenous self-determination.

(d) Ethics: Indigenous rights and wellbeing must be the primary focus across the entire data life cycle to minimize harm and promote justice.

The above provides the ultimate guiding reference for the pilot’s aim to build and sustain a dashboard where the Ngäbe Council of Elders can technically enforce whether their data is public, restricted, or sacred. This is real veto power.

Secondly, the pilot will need to consistently prove that memory systems are able to work where internet connectivity is virtually nonexistent. The pre-condition of this lies in the development of a PWA (Progressive Web App) in order to sync peer-to-peer over mesh networks. It has to run on older Android phones with under 500MB of RAM. In essence, if it works there, it works anywhere.

Third, co-design a simple, culturally-sensitive and language-appropriate NLP model for Ngäbere. Available figures indicate nearly 200,000 people speak it, but it has almost no digital footprint. Hemispher-IA will hence pursue the release of an open dataset, a basic transcription pipeline, and a playbook. With such pipeline proving functional, adapting it at a later stage to K'iche' in Guatemala or Guaraní in Paraguay will cost almost nothing (And the pilot’s intention, if succesful, is to become a replicable, scalable model in the entire Latin America and the Caribbean region.

We achieve this by combining academic rigor, with direct advice from Recordari’s creators and Universidad del Rosario, strict ethical auditing via CLET and CAIDP, and by actually paying local Indigenous talent to co-design and maintain the system.

How will this funding be used?

The requested $150,000 is meant to cover 18 months of working capital, detailed as per the below estimations:

About 68 percent goes to human capital. Indeed, we are completely against volunteer models that burn people out. The requested amount pays for executive direction, a senior technical lead, binational field coordinators, in addition to six Indigenous Digital Weavers at $450 a month plus tech kits. Believe it or not, in the indigenous comarcas, a formal contract with benefits is often the difference between a young person staying in their territory or migrating to Panama City to work in a call center.

Roughly 15 percent funds tech. Local servers, PWA development, and the NLP fine-tuning pipeline.

Another 7 percent is for field equipment. Refurbished smartphones, solar power banks, mesh antennas.

The remaining 10 percent handles legal advisory for the binational complexity between Panama and Costa Rica, the CLET ethical audit, and a small operational contingency.

Field deployments always hit unexpected hurdles. This buffer accounts for that.

Who is on your team? What's your track record on similar projects?

Hemispher-IA is a Panamanian social innovation initiative. It focuses on high-level knowledge brokering and managing scalable portfolios across Latin America and the Caribbean.

Executive direction is led by Wolfgang Friedl. He has over a three decades of experience building institutional partnerships for social innovation across the region, including work with UNICEF, UNIC, UNODC, WHO, and diplomatic organizations. His expertise lies in assembling the right technical and community partners and ensuring projects do not collapse under their own administrative weight.

The technical aspect and the ideological backbone comes from advisors Arthur Picerna and Luisa María Rincón at Universidad del Rosario, who co-created Recordari; an open-source architecture for resilient historical memory in post-conflict Colombia. They will now undertake the proper adaptations of those offline-first data structures for an Indigenous context.

On the ground, OLINABP, the Ngäbe-Buglé Future Leaders Organization in Panama, and FAIR Costa Rica will serve as the primordial operators. They are Indigenous-led and local responsible-AI organizations with direct legitimacy and a proven track record in the pilot regions.

The governance board includes Latin America Council on Ethics in Technology, CLET for ethical audit and CARE certification, Center for AI & Digital Policy, CAIDP for human rights and privacy, Global Council for Responsible AI, GCRAI for global AI alignment, the Foundation for Responsible AI, FAIR Costa Rica for local adaptations and review of relevant regulations, and the Silicon Valley Certification Hub for auditable AI standards. 

The track record relies on validated partnerships and deployed frameworks, not a long list of academic papers. The Recordari architecture works in the field. The CLET audit process is established. Combining all three in a binational Indigenous context is the specific gap this pilot addresses.

What are the most likely causes and outcomes if this project fails?

Integrating Indigenous epistemologies into global AI infrastructure is a multi-decade challenge. It is highly likely we will not achieve complete continental adoption within the initial 18-month timeline. Legal fragmentation between Panama and Costa Rica alone could cause significant delays. Community adoption might be slower than we hope. The NLP work might yield a modest dataset rather than a robust model.

But even if the broader vision stalls, the funding will still produce high-leverage, open-source public goods. The CARE implementation code will be on GitHub for anyone to fork. The offline-first PWA will function in any low-connectivity environment. The NLP pipeline for Ngäbere will serve as a methodological template for dozens of other under-resourced languages. Lastly, the binational replication playbook will openly document the technical, ethical, and social bottlenecks we encountered. This will save the next team months of trial and error. We prefer building durable public goods that survive a partial failure over a flashy demo that disappears when the grant ends.

How much money have you raised in the last 12 months, and from where?

Over the past 12 months, the initiative has secured significant in-kind contributions and seed support. This includes technical advisory hours from the Recordari team, a committed pro-bono ethical audit from CLET, and formal community access plus field infrastructure from OLINABP and FAIR Costa Rica. We are currently in preliminary conversations with institutional partners like Senacyt in Panama and KAS-PPI regarding co-funding. This Manifund grant would represent the first targeted cash injection specifically for the Memoria Viva Ngäbe pilot. It is necessary to transition from planning and in-kind support to active development, hiring, and field deployment.