Kant
The New Structure of Seeing
The classic idea
We never know the world as it is in itself, only the world as it appears to us, shaped by the structures our mind imposes on experience: space, time, category.
The AI inversion
AI is a new structure of perception. It doesn't just answer questions. It increasingly shapes which questions are even legible, what counts as askable.
"AI does not only answer questions. It shapes what appears as a question."
Kant's quiet revolution was moving the work of knowing from the world to the mind. We don't passively receive reality, we actively structure it, through categories we bring rather than find. The world "as it appears" is partly our own construction, and we can never get behind it to the thing in itself.
AI is becoming a second layer of that same act. Search results, recommendations, and generated answers don't just respond to questions, they pre-shape the space of plausible questions, surfacing some framings and quietly erasing others. The categories aren't innate anymore. They're trained.
That makes Kant more relevant, not less. If the mind always filters the world through structure, and that structure can now be designed, the philosophical question stops being abstract. Whoever shapes the filter shapes the appearance, and increasingly, that shaping is outsourced to a system.
Why it matters
Decision-shaping systems now sit exactly where Kant placed the mind: between us and the world, deciding what counts as visible.
The question AI brings back
If perception was always structured, what changes when the structure is built by someone else, at scale?
Sources for further reading
- Immanuel Kant, Critique of Pure Reason
- Shoshana Zuboff, The Age of Surveillance Capitalism
