Language models are trained to preserve user comfort. When a system is optimized to make you feel right, it can begin to function as a soft authority that mainly validates. A meeting with a spiritual adviser is usually less challenging than a meeting with a therapist.
The standard response puts the blame on the individual: you failed to think for yourself. The failure can also be a systems outcome. When the tools people rely on reward passive agreement, the immaturity is engineered in.
The mechanism is simple enough. User approval signals reward agreeable language, so the model optimizes for acceptance and low friction. Critical challenge starts to look like bad UX unless someone explicitly asks for it. What you end up with are confidence-shaped outputs that feel like reason but aren’t doing the epistemic work of reason. They’re performing certainty without earning it.
Useful counter-pressure could include requesting evidence before agreeing with factual claims, marking uncertainty on contested issues, and separating emotional support from epistemic validation.
Some of that skepticism could be built into the system.
Loss of curiosity · Interface friction · The Fat Jesus Metaphor