# The Effects of Interaction Scenarios on EFL Learners’ Technology Acceptance and Willingness to Communicate with AI

## TL;DR

This study explores how different AI interaction scenarios affect Chinese EFL students' acceptance of technology and their willingness to communicate with AI.

## Contribution

The study introduces a scenario-sensitive framework that extends TAM and WTC theories in AI-mediated language learning.

## Key findings

- Factual information retrieval scenarios received highest ratings for perceived usefulness and ease of use.
- Advisory interaction scenarios were perceived as more useful than language skills support scenarios.
- Low willingness to communicate emerged in complex scenarios despite perceived usefulness.

## Abstract

Grounded in a sociocultural theory, this study investigates how distinct interaction scenarios influence Chinese English as a Foreign Language (EFL) learners’ technology acceptance: perceived usefulness (PU) and perceived ease of use (PEU), and their willingness to communicate with AI (AI-WTC). A total of 367 university students completed a scenario-based questionnaire measuring PU, PEU, and AI-WTC across four empirically derived scenarios: advisory interaction, language skills support, academic knowledge inquiry, and factual information retrieval. Repeated-measures ANOVA with Bonferroni tests revealed significant scenario effects on all three constructs, though effect sizes were small to moderate. Factual Information Retrieval Scenario consistently received the highest ratings, whereas Academic Knowledge Inquiry and Language Skills Support Scenario scored lowest. A salient divergence emerged in complex scenarios: Advisory Interaction Scenario was rated more useful than Language Skills Support Scenario, yet both elicited equally low willingness to communicate, indicating that perceived usefulness alone may not sustain engagement under high interactional demands. These findings suggest that the effectiveness of AI as a communicative scaffold is not inherent but co-constructed through scenario-specific affordances and constraints. The study contributes a scenario-sensitive framework to TAM and WTC research, providing pedagogical guidance for designing differentiated AI-mediated language tasks.