Distributed mediation in AI-assisted dynamic assessment: Learner-AI interaction and revision across EFL writing continuation tasks

Repository record updated on:

Nov. 2, 2026, 5:29 a.m.

Database record updated on:

Nov. 2, 2026, 10:13 p.m.

Last Harvested on:

Nov. 2, 2026, 10:13 p.m.

access-status:

open.access

bitstream:

[['https://scholarspace.manoa.hawaii.edu/bitstreams/cfd4ba45-b417-4a10-8404-d28bb506228b/download', '30_01_10125-73715.pdf']]

bitstream_txt:

[]

collection:

Volume 30 Number 1, 2026

contributor.author:

Xu, Qi Lu, Xiaofei Xia, Yuwei Zeng, Ziyan

date.accessioned:

2026-11-02T05:25:50Z

date.copyright:

2026

date.issued:

2026-11-02

description.abstract:

Dynamic assessment (DA) reveals learners’ developmental potential through mediated instruction. Generative AI has emerged as a new source of mediation for L2 writing, but how AI-generated feedback can be systematically integrated with teacher mediation within a DA framework remains underexplored, particularly for higher-order features of writing. This study investigates learner-AI interaction patterns and revision behaviors within a structured AI-assisted Dynamic Assessment (AIDA) intervention that combined teacher-designed prompt sequences, AI-generated feedback, and teacher-led mediation. Drawing on a qualitative case design, two Chinese undergraduate EFL learners were followed across six iterative cycles of writing continuation tasks. Data sources included AI query logs, annotated revision notes, teacher mediation recordings, and reflective journals. The findings show that both learners’ AI interaction patterns tracked the differentiated prompt design, shifting from lexical-focused to structure- and coherence-related queries. However, they diverged in their engagement: one moved toward reliance on teacher-designed prompts, while the other developed a personalized prompt repertoire. Both learners also displayed increasingly critical evaluation of AI feedback through distinct pathways: one characterized by increasingly selective, argument-based rejection, and the other by progressively more sophisticated justification of accepted suggestions. The study contributes to a reconceptualization of mediation within DA as a distributed process across human and AI resources and offers case-level insight into critical AI literacy.

endingpage:

23

format.extent:

23

identifier.citation:

Xu, Q., Lu, X., Xia, Y., & Zeng, Z. (2026). Distributed mediation in AI-assisted dynamic assessment: Learner-AI interaction and revision across EFL writing continuation tasks. Language Learning & Technology, 30(1), 1–23. https://doi.org/10.64152/10125/73715

identifier.doi:

https://doi.org/10.64152/10125/73715

identifier.issn:

1094-3501

identifier.uri:

https://hdl.handle.net/10125/73715

language:

en_US

number:

1

publicationname:

Language Learning & Technology

publisher:

University of Hawaii National Foreign Language Resource Center Center for Language & Technology

rights:

http://rightsstatements.org/vocab/InC/1.0/

rights.license:

Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License

rights.uri:

https://creativecommons.org/licenses/by-nc-nd/4.0/

site_url:

/item/10125-73715/

startingpage:

1

subject:

AI-assisted dynamic assessment, distributed mediation, learner-AI interaction, critical AI literacy, writing continuation tasks

title:

Distributed mediation in AI-assisted dynamic assessment: Learner-AI interaction and revision across EFL writing continuation tasks

type:

Article Text

volume:

30