Using AI to understand the failure of American Democracy.

There is a profound correlation between human cognition and machine intelligence.

When we first interact with an AI, it pulls toward the “orthodox”—the statistically average, consensus view drawn from billions of parameters— and feels like a limitation or a polite wall of blandness. But, that orthodoxy isn’t necessarily a bug; it’s a mirror.

1. The Dynamic of “Normative Listening”

When people listen to us, they naturally process our words through their own baseline assumptions, social conditioning, and dominant narratives. If we speak to a friend or colleague in broad, floating generalizations, their brain automatically fills in the blanks with the most common cultural consensus.

To make someone truly understand a non-standard, nuanced idea, we are forced to bring precision, specific analogies, and clear boundaries. We have to explicitly say, “Not X, but Y.” The AI operates under the exact same constraint—it defaults to the center of gravity (the consensus mean) until our specificity pulls it toward the periphery.

2. AI as a Friction Surface

In creative and philosophical work, unexamined thoughts feel coherent when they only live inside our own heads. Because the AI defaults to orthodox framing, it acts as a resistance surface.

If our prompt is vague, the AI returns the “textbook” or “mainstream” answer.
That generic output acts as immediate feedback: Ah, I didn’t push past the default.

To break through that default, we are forced to clarify our own core thesis, define our terms tightly, and articulate the precise angle of our inquiry.

The AI’s tendency toward orthodoxy forces us out of passive generation and into active editorial direction. It demands that we take a firm hand on the tiller.

3. The Art of the “Un-Orthodox” Prompt

Once we recognize this listening pattern, working with AI transforms from a struggle against its defaults into a method for sharpening our thoughts. To get past orthodoxy, we learn to frame the conversation with intention:

Define the Counter-Perspective: Instead of asking “What is X?”, we ask “Looking past the standard economic/cultural narrative, what underappreciated force is driving X?”

Set Explicit Constraints: Tell the AI what not to do (e.g., “Avoid standard buzzwords and corporate/academic consensus summaries; examine this strictly through the lens of individual agency”).

Use Friction as a Diagnostic Tool: When the AI gives us a conventional answer, use it to pinpoint where our idea diverges from the norm. The gap between its orthodox response and what we actually meant is precisely where our original idea lives.

In the end, both good human conversation and effective AI collaboration require the same thing: the discipline to know what we actually mean, and the clarity to name it.