Turn-taking with an AI partner shows slower responses, reduced learner floor share, and greater within-turn fluency¶
Claim · All claims
Evidence · 1 study · causalr2·q3·i3large
Currency · newest study 2026 · not yet checked against its sources · not reviewed by a person
Subclaims¶
Reading the evidence codes
q is the design tier, i the impact, n the sample, and the kind with its r
the rigour judged within that kind: what each code means.
This claim: no standardized effect is recorded in observations/ for it yet, so no benchmark reading is possible.
q3 i3 Participants took roughly twice as long to begin speaking after the AI (d=0.90), held the floor for a smaller share of session time (d=−1.14), and paused less within their own turns (d=−0.53) than with a human partner. → Russell Scheinberg 2026
Evidence¶
Russell Scheinberg 2026¶
Russell Scheinberg, Ameeta Agrawal, Tetyana Sydorenko, Kalab Kahsay, Nina Vyatkina, Griet Boone. (2026). What Changes When the Interlocutor Is an AI? Interactional Fluency and Linguistic Uptake in L2 Spoken Dialogue. https://arxiv.org/abs/2606.22225 doi:10.48550/arxiv.2606.22225
q3 · i3 · causal · r2
Turn-taking analysis of sessions with at least 8 participant turns from the same within-participant study. Response latency (d=0.90), floor ratio (M=0.37 vs 0.48, d=−1.14, "representing one of the largest effects observed in the dataset"), and within-turn pause ratio (d=−0.53) all differed significantly (p<.001).
"Participants took roughly twice as long to begin speaking after the AI (M= 0.61 s) as after a human partner (M= 0.32 s;d= 0.90,p < .001)"
Discussion¶
Related Claims¶
- Human-human L2 dialogue has more and shorter turns while AI dialogue has fewer, longer turns with more words per turn — related
- Attitudes toward AI improved after direct experience, and satisfaction was predicted by within-turn production fluency rather than linguistic uptake — related
- Despite longer AI-led sessions, the proportion of student-generated dialogue was markedly and significantly lower with the AI than with the teacher in both grades — related
- In individual dialogue patterns, the dominant speaker learns more than the listening partner, potentially explaining null collaboration findings — related
- Interactive engagement requires both partners' utterances to be primarily constructive and a sufficient degree of turn taking — related
- Voice modality roughly doubled dialogue turns and question-asking on equal engaged time within the same students — related
- Teaching interactions typically conclude within 30 dialogue turns, motivating a 30-turn cap to prevent dead loops — related
Page metadata
| Field | Value |
|---|---|
type |
claim |
id |
ai-turn-taking-latency-floor-pause |
title |
Turn-taking with an AI partner shows slower responses, reduced learner floor share, and greater within-turn fluency |
description |
Turn-taking with an AI partner shows slower responses, reduced learner floor share, and greater within-turn fluency |
status |
review |
evidence_strength |
moderate |
newest_evidence |
2026 |
generated |
by process:wiki-ingest on 2026-10-10 |
trust |
unverified — no one has confirmed this page's content |
sources |
1 — russell-scheinberg-2026 |
Frontmatter is the page's machine-readable record — see the Schema & Guide.