Attitudes toward AI improved after direct experience, and satisfaction was predicted by within-turn production fluency rather than linguistic uptake¶

Claim · All claims
Evidence · 1 study (2 entries) · causal r2 · q3 · i2 medium
Currency · newest study 2026 · not yet checked against its sources · not reviewed by a person

Subclaims¶

q3 i2 Directional sentiment scores increased significantly after the task (mean +0.48, d=0.64), with 75% of participants becoming more positive. → Russell Scheinberg 2026 q3 i? Within-turn pause ratio was the strongest behavioral predictor of satisfaction (β=−4.78, p=.008); neither lexical adoption rate nor syntactic uptake rate predicted satisfaction (p>.45). → Russell Scheinberg 2026 (2)

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 · i2 · causal · r2

Paired t-test on pre/post surveys from 55 matched respondents (52 with complete composite data) in the multi-site study. Post-study, 67.3% liked the AI interaction and 81.8% agreed AI can be useful for L2 speaking practice.

"a pairedt-test on directional sentiment scores showed a mean increase of +0.48 (t(51) =−4.58,p<.0001, d= 0.64), with 75% of participants becoming more positive."

Russell Scheinberg 2026 (2)¶

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

q3 · i? · causal · r2

OLS regressions predicting post-satisfaction and attitude change from AI-dialogue behavioral features (n=55, site dummy-coded). No standardised effect size is printed for these regression coefficients. Floor ratio was not significant, suggesting how fluently learners spoke mattered more than how much they spoke.

"the strongest behavioral predictor was within-turn pause ratio (β=−4.78,p=.008): learners who paused more during AI turns, reflecting greater online planning difficulty, reported lower satisfaction and smaller attitude gains (β=−5.91,p=.006)."

Discussion¶

Related Claims¶

• Turn-taking with an AI partner shows slower responses, reduced learner floor share, and greater within-turn fluency — related
• Learners show stronger short-term syntactic priming with the AI partner, significant across all three recency-window operationalizations including one equalizing input volume — related
• Public sentiment toward AI in education on Twitter was predominantly positive (81.65% of tweets) across 2019–2024 — related
• Interactive dialogue with the AI tool, available only in some sites, promoted deeper engagement and student agency during revision — related
• Keystroke patterns during web-based DP exams suggested planning, revision, and sustained cognitive engagement — related
• Hands-on engagement with an AIDA beta significantly improved students' attitudes toward it, with a large effect size — a narrower finding that bears on this claim
• After hands-on use, 96% of students wanted AIDA in their studies, with the quiz function the most liked feature — related
• Lexical adoption shows no significant condition effect after controlling for input volume, but AI-first participants adopted more words at the second dialogue position — related