Prompting Pattern of MAN 2 Malang City Students in Writing ChatGPT-Assisted Reviews

DOI: https://doi.org/10.26618/ngge6963

Penulis

  • Qolbi Azizatul Universitas Negeri Malang
  • Moch. Syahri Universitas Negeri Malang
  • Gatut Susanto Universitas Negeri Malang

Abstrak

 

This study aims to explore the prompting pattern of students in interaction with ChatGPT in review text writing activities. The main focus of the research lies in how students design and use prompts as strategies in directing AI responses at the revision stage (rewriting), not solely on the results of the writing produced. This study uses a qualitative approach with a descriptive-analytical design. The research subjects consist of students in two Phase F classes specializing in Science and Technology at MAN 2 Malang City. The learning activity adopts Donald M. Murray's writing process strategy so that ChatGPT is specifically integrated as a dialogical partner in the revision phase of review texts. Data in the form of 154 prompts obtained from 31 units of student interaction with ChatGPT which were analyzed through a thematic coding process.

The results showed that student interaction formed a diverse pattern of prompting strategies, including evaluative prompting, directed revision, and response control. These strategies reflected different levels of cognitive involvement in utilizing AI as digital scaffolding in students' Zone of Proximal Development (ZPD). However, most prompts were still general and technical improvement-oriented, indicating efforts to minimize cognitive load in the revision process. These findings show that the quality of interaction with AI is highly dependent on students' ability to design prompts specifically and reflectively. This study emphasizes that prompting is a key competency in AI-based learning and needs to be explicitly developed as part of digital literacy in pedagogical practice.

Keywords: ChatGPT, AI literacy, digital learning, prompting, review

Diterbitkan

2026-06-29

Terbitan

Bagian

Artikel