This study examined Indonesian EFL learners' perceptions and attitudes toward ChatGPT in English language writing and the relationship between them, a pairing few studies have addressed simultaneously. Using a quantitative descriptive-correlational design, data were collected from 110 Indonesian EFL students through an 18-item Likert-scale questionnaire (Cronbach's α = 0.894). Students held positive perceptions of ChatGPT as a writing support tool (M = 3.75), while overall attitudes were neutral (M = 3.22). This neutral aggregate, however, conceals a meaningful internal structure: skill degradation concern was the dominant attitudinal response (A4: M = 4.05, 78.2%), while full AI delegation and uncritical output adoption were the most strongly rejected behaviors (A8: M = 2.86; A9: M = 2.68), suggesting an “aware but resisting passive dependence” orientation. A moderate positive correlation between the two constructs (r = 0.482, p < 0.001) confirms that utility appraisals shape, but do not fully determine, attitudinal formation. These findings suggest that learners are already exercising evaluative agency over AI use, and that instruction should build on rather than ignore this disposition.
Copyrights © 2026