As artificial intelligence (AI) becomes increasingly embedded in society, it's essential to pause and reflect on the foundations that sustain it—and the dimensions to which it extends. At the heart of AI's learning are datasets, whose structure and content shape how these systems interpret and respond to the world. This reliance creates a deep interdependence—one that not only informs AI's capabilities but also defines its potential blind spots. In light of this, we must ask: What forms of understanding might this process exclude, especially those not easily captured in digital form?
