DOI: 10.4103/picr.picr_193_26 ISSN: 2229-3485

Practice patterns, determinants, and outcomes of artificial intelligence use in research writing among postgraduate health science students in South Karnataka, India

Uma Kulkarni, Akuma Ifeanyichukwu, Aarti Halwai, Khadeejath Farseena, Shama U. Rao, Naheeda Hamza, Josna B. R. John
Show PDF Cite

Abstract

Background:

Artificial intelligence (AI) technology is increasingly supporting research writing, raising concerns about academic integrity, intellectual skill development, and ethics in academics and research.

Methodology:

A cross-sectional survey was conducted from February to May 2026 among 383 postgraduate students from four randomly selected health science institutions in Mangalore, Karnataka, to assess AI usage, determinants of utilization, and perceived outcomes. Descriptive statistics were used to summarize the data, while Chi-square tests and multivariable logistic regression analyses were performed to examine associations and identify factors associated with AI utilization.

Results:

Majority of respondents reported familiarity with ChatGPT (98.2%) and previous AI usage (68.7%). Age, program of study, research experience, and prior AI usage were statistically associated with AI utilization (all P < 0.05). AI application was mostly to paraphrase (62.8%), summarize (61.9%), and check grammar (57.4%). Major facilitators included perceived usefulness (mean [ M ] = 3.82, ±standard deviation: ±0.91) and ease-of-use ( M = 3.91 ± 0.83). Among the facilitators assessed, institutional norms (perceived institutional support and expectations regarding AI use) ( M = 3.01 ± 1.01) and confidence in AI output ( M = 2.98 ± 0.895) had the lowest mean scores. The highest-rated perceived benefit of AI use was improved time efficiency gained ( M = 3.95 ± 0.83), followed by enhanced academic performance ( M = 3.66 ± 0.84). Concern about plagiarism received the highest concern rating ( M = 3.99 ± 0.80). Regression analysis identified ethical awareness as the only statistically significant independent predictor of AI use (β = 0.100, P = 0.019).

Conclusion:

Postgraduate students widely adopt AI tools to support academic and research writing, particularly to improve writing efficiency. However, AI utilization appears to be influenced more by students’ ethical awareness than by conventional demographic or academic factors. Responsible and effective integration of AI into postgraduate education and research within health science institutions requires clear institutional AI in research guidelines that promote ethical, transparent, and accountable AI use.

More from our Archive