From Silence to Simulation: Exploring AI as a Safe Space for Negotiating Speaking Anxiety among ESP Students

Abstrak

Speaking anxiety remains one of the most significant affective barriers in English for Specific Purposes (ESP), particularly for engineering students who are expected to communicate using both general English and discipline-specific terminology. Although artificial intelligence (AI) has increasingly been integrated into language learning, limited qualitative research has examined how AI-mediated speaking practice helps learners negotiate speaking anxiety within ESP settings. Therefore, this study explores how AI-based speaking practice, particularly through the use of Stimuler AI, functions as a safe space for negotiating speaking anxiety among civil engineering students. This study employed a qualitative case study design involving 32 second-semester students in the Road and Bridge Construction Study Program. Data were collected through semi-structured interviews, reflective journals, and AI interaction logs and analyzed using thematic analysis. The findings revealed that students initially experienced high levels of speaking anxiety characterized by fear of negative evaluation, avoidance of participation, and concerns about using technical English. Through repeated interaction with Stimuler AI, students gradually developed greater confidence, improved their technical communication skills, and became more willing to participate in classroom speaking activities. The study also identified a transformation from classroom silence to active engagement and emerging professional identity as future engineers. These findings suggest that Stimuler AI provides a psychologically safe environment that enables learners to negotiate speaking anxiety while supporting speaking development and identity construction in ESP classrooms.

Identifier

• OpenAlex: https://openalex.org/W7171800692
• DOI: https://doi.org/10.26877/allure.v6i2.524