Tuesday, October 13, 2026

  • Artificial intelligence isn't transforming education. It's exposing it. The most dangerous assumption shaping the future of education isn't that AI is biased.

    It's that education was ever neutral.

    For decades, we've mistaken compliance for learning, efficiency for excellence, standardization for rigor, and data for understanding. Artificial intelligence didn't invent those assumptions.

    It inherited them.

    Now it's teaching them back to us as if they were objective truth.

    But what if the limitations of AI reveal less about technology than they do about us?

    Large language models were built on assumptions rooted in Western languages and Western knowledge systems. Around the world, many languages encode relationships, context, and meaning in ways that those systems were never designed to represent.

    Research in Indigenous language revitalization reveals that this is not simply a translation problem. It is evidence that language itself carries fundamentally different ways of organizing knowledge.

    Drawing from her book, research, and teaching at the intersection of AI, statistics, and Indigenous knowledge systems, Karen Colbert reveals the invisible architecture beneath modern education and the assumptions artificial intelligence is now scaling at unprecedented speed.

    Introducing a novel educational framework for the age of artificial intelligence, this keynote offers a new scientific lens for evaluating intelligence, learning, and educational design.