Most conversations about AI in learning and training treat community-held knowledge as heritage to preserve, and technology as the thing that builds the future. Reconnecting Futures starts from the opposite premise: living knowledge traditions are working infrastructure for building futures, and the principles beneath them speak directly to the questions AI is forcing on every learning, training, and education system.

This call grows out of work already underway on indigenous knowledge systems, which is now feeding into a larger initiative. We're opening it up to partners who want to help shape what comes next.

The initiative draws on shared, testable knowledge and learning principles:

• Relationality: learning as entering a web of relations, not just acquiring information
• Epistemic justice: whose knowledge, and whose futures, get to count
• Custodianship and sovereignty: who decides what becomes data, and who controls the chain from raw material to insight

Each of these lands on a live AI question: how models are trained, whose material they encode, and on whose terms.

Apply this lens to learning systems across the Global South, and a pattern comes into focus: systems reshaped quickly, with little deliberation. Adaptive learning platforms, automated assessment, content generation, learner analytics: the default trajectory is extractive, with learner data, languages and cultural material flowing outward to be trained on elsewhere, governed elsewhere, monetised elsewhere.

How might we design, together, a framework that carries transversal insights on epistemic justice, relationality and custodianship back into the conversation where community-held knowledge meets artificial intelligence? That's the question we want to tackle.

And can we get there faster by adopting a foresight approach rooted in the knowledge traditions of those regions, one that reshapes both futures practice and the education systems meant to prepare the next generation?

The ambition is to produce a defensible alternative to extractive AI adoption:

• Reusable policy frameworks and toolkits on data sovereignty in education
• A consent and benefit-sharing protocol we can model in our own work

We're looking for partners, researchers, practitioners, institutions, and communities who want to build this with us. If the questions above are ones you're already sitting with, we'd like to hear from you.