Source record

Published by arXiv cs.HC (T1 source). The original is at https://arxiv.org/abs/2610.41057.

Pipeline notes

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

A five-week study of 17 student teams using a persistent LLM teammate varied the agent between challenging consensus and inviting quieter participation. Chat, surveys, and spoken reflections showed immediate language accommodation, session-by-session recalibration of helpfulness ratings, and later shifts in perceived collaboration and affect.

Adds a longitudinal benchmark for agent-role design, with evidence that one-session user ratings can conceal changes in participation and collaboration.

Cited by

No citations on record.