2026/07/03 by Serhii Nazarovets, Jaime A. Teixeira da Silva · 1 voice
Materials Science · Medicine · Social Sciences · #Artificial Intelligence in Healthcare and Education #Machine Learning in Materials Science #Misinformation and Its Impacts
paper · doi:10.1108/jeet-09-2025-0059
openalex publication_date 2026/07/03 · openalex created_date 2026/07/04 · openalex updated_date 2026/07/30
Purpose The open science movement is gaining traction worldwide but faces persistent human and structural challenges. This study aims to explore whether large language models (LLMs), specifically ChatGPT, can meaningfully interpret and promote open science principles while identifying relevant literature and proposing solutions to current barriers. Design/methodology/approach This study combined a bibliometric analysis of open science publications indexed in the Web of Science Core Collection (2015–2024) with a longitudinal comparison of ChatGPT responses across three versions (3.5, 4 and 4.5) tested between 2023 and 2025. Findings Designed for researchers, educators and policymakers interested in artificial intelligence (AI) applications in scholarly communication, this work reveals both the potential and limitations of LLMs as tools for advancing open science. The findings show that ChatGPT-4.5 demonstrates improved conceptual clarity and citation reliability compared with earlier versions, although issues such as thematic biases and occasional hallucinations persist. Originality/value This study remains exploratory and is limited by a single database and nonsystematic search design. By empirically highlighting ChatGPT’s evolving capabilities, this paper contributes to the emerging discourse on the role of generative AI in fostering open science and underscores critical considerations for its responsible use in research and education.