Dependence on the Use of ChatGPT and Its Impact on Statistics Students' Linguistic Literacy in Compiling Data Analysis Reports

Abstrak

The rapid development of generative artificial intelligence, particularly ChatGPT, has significantly influenced higher education, especially in student academic writing activities. However, research specifically examining the relationship between ChatGPT dependency and student language literacy in preparing data analysis reports remains limited. This study aimed to analyze the relationship between dependency on ChatGPT usage and the language literacy of Statistics students in preparing data analysis reports. A quantitative approach with a survey strategy was employed. The population consisted of 265 active Statistics students at Universitas Negeri Medan from the 2023, 2024, and 2025 cohorts, with 66 respondents selected through purposive sampling. Data were collected using a digital questionnaire via Google Forms, measuring ChatGPT dependency (7 items) and language literacy (6 items) on a 4-point Likert scale. Descriptive statistical analysis and Pearson correlation tests were conducted using RStudio. The results showed that ChatGPT dependency was in the Agree category (mean = 2.70), with students demonstrating selective use by disagreeing with directly copying AI output (1.91). Language literacy was also in the Agree category (mean = 2.98), although confidence in independent writing remained low (2.62). The Pearson correlation test revealed no significant relationship between ChatGPT dependency and language literacy (r = -0.002; p = 0.9903 > 0.05). The study concludes that AI dependency does not inherently weaken language literacy, and educational efforts should focus on critical AI usage rather than prohibition.

Keywords: ChatGPT dependency, language literacy, generative artificial intelligence, academic writing, Statistics students.