Governing artificial intelligence in science: Evidence from a bibliometric analysis of research integrity

The increasing integration of artificial intelligence into scientific research has raised worries about scientific ethics and research integrity. This research offers a thorough bibliometric analysis. Of 2969 publications Scopus index from 2015 until 2025 to examine the field's evolution. The findings reveal a rapid expansion of research output, with annual publications increasing from 19 in 2015 to 1143 in 2025 and an annual growth rate of 50.64 percent. The United States leads in productivity, with 715 publications and 21336 citations, while the United Kingdom demonstrates the highest citation impact, with a citation-per-paper value of 50.75. A total of 12077 authors contributed to the field, reflecting strong collaboration patterns, with 28.83 percent of publications involving international co-authorship. Thematic analysis identifies five major research clusters: ethics, artificial intelligence technologies, methodological research, evaluation processes, and scientific communication. These results highlight a rapidly evolving, interdisciplinary domain characterized by growing scholarly attention and complex ethical challenges.