Diana Indah Yuniarti. The Use of Artificial Intelligence for Scientific Writing from Socio Cognitive Theory: An Interview Study. Thesis. Department of English Language Education, Muhammadiyah University of Ponorogo. Advisor 1 Dr. Elok Putri Nimasari, M.Pd., Advisor 2 Dwiana Binti Yulianti, M.Pd.

This research aims to: 1. describe how students from different study programs perceive the role of artificial intelligence in supporting their scientific writing, 2. explain how the use of artificial intelligence affects students' self-efficacy, observational learning, and self-regulation in scientific writing based on Bandura's Socio Cognitive Theory, and 3. analyze the interaction among personal, behavioral, and environmental factors when students use artificial intelligence for scientific writing.

This research is a qualitative study using a semi-structured interview approach and data artifacts, conducted at the Faculty of Teacher Training and Education, Universitas Muhammadiyah Ponorogo. The participants consisted of ten sixth-semester students who have used AI tools for academic writing, selected through purposive sampling from four study programs, namely English Language Education, Mathematics Education, Civics Education, and Early Childhood Teacher Education. The data collection technique used was face-to-face, semi-structured interviews guided by open-ended questions developed from the constructs of Socio Cognitive Theory, while the instrument used to collect the data was an interview guide supported by a voice recorder. The data analysis technique used was thematic analysis as outlined by Braun & Clarke (2006), and a deductive approach in which Bandura's Socio Cognitive Theory (2001).

The results show that all participants positioned artificial intelligence as a supplementary writing tool rather than an independent author, functioning as a co-assistant, structural scaffold, language enhancer, and reference aid. The use of artificial intelligence generally strengthened students' self-efficacy through mastery and vicarious experiences, supported observational learning through abstract modeling, and prompted active self-regulation, although this regulation weakened under time pressure. The interaction among personal, behavioral, and environmental factors followed a reciprocal pattern consistent with Bandura's triadic reciprocal determinism, indicating that artificial intelligence-assisted writing functions as a continuously evolving system shaped jointly by the writer, writing behavior, and the surrounding academic environment, rather than as a one-directional technological influence.