Source record

Published by arXiv cs.HC (T1 source). The original is at https://arxiv.org/abs/2607.17627.

Pipeline notes

The summary and note below are generated by the signal pipeline — they are Beyond Desk’s reading, not quotations from the source.

Study with 98 participants testing eight rhetorical patterns for AI-assisted fact verification, including Socratic Questioning, Alternative Framing, and adversarial styles. Scaffold Explanation produced the highest accuracy gains; participants preferred Alternative Framing but found Interpretive Alternative too time-consuming.

Quantifies how the rhetorical packaging of AI responses changes verification accuracy and user reflection, separating stated preference from measured performance. Useful before designing any agent that nudges users toward critical evaluation.

Cited by

No citations on record.