Abstract Artificial Intelligence (AI) tools are used in the academic research, clinical practice and education in medicine. However, few is known on the impact of AI tools used on publication performance of researchers. Purpose: The aim of our study was to explore the level of AI tools use among physicians and relevant disciplines involved in medical research and assess their impact on publication performance. Methods: Anonymous survey study was conducted according to COREQ guidelines. The questionnaire was prepared and distributed among physicians and relevant disciplines involves in medical research using Google survey form, or Adobe document, ‘snow-balling’ is allowed. For statistical analysis we used descriptive statistics and parametric paired t-test or nonparametric Friedman test, according to normality of distribution, were used for comparison of continuous variables before and after AI use. Results: 71 people agreed to participate, with a final group of 51 researchers from 10 countries (Australia, Brazil, Cyprus, Italy, Kazakhstan, Kyrgyzstan, Poland, Russia, Turkey, Ukraine). The participants mean age was 41.7±9.8 years, gender: male 22/43.1%, female 29/54.9%, prefer not to say 1/1.96%. Place of work: University – 19/37.2%, Academic/Teaching hospital – 21/41.17%, Local hospital -6/11.76, Private practice 2/3.92%, Other 2/3.92%, Scientific institution 1/1.96%. Position: Senior researcher /DSc/Senior faculty/full professor – 10/19.6% , PHD degree holder/PostDoc researcher/Junior faculty – 17/33.3%, PhD students – 14/27.4% , Fellows in training /Residents - 5/9.8% Intern - 1/2%, Other - 3/5.9%. Among 51 respondents 43/84.3% use AI and 8/15.7% did not, 33 knew the time for article preparation and 16 had Hirsch index or acquired Hirsch index after AI use. They mainly used: ChatGPT - 35/68.6%, DeepSeek - 12/23.5%, Grammarly - 13/25.4%, Humanize AI - 3/5.88%, other - 6/11.7%. (Fig. 1) Comparison of reported number of publications in indexed journals (SCOPUS, Clarivate, Pubmed) before and after AI use (Table 1) showed significant jncrease in publications after AI use (p 0.001) and increase in Hirsch index (p=0.017). Conclusions: Our preliminary data showed that almost 71% of surveyed physicians use AI. Among AI tools ChatGPT is used most frequently by almost 76%. The use of AI tools increased quality of publications and academic performance , i.e. number of articles published in indexed SCOPUS, CLARIVATE, PUBMED, MEDLINE journals and increased Hirsch index. The time for preparation of article for publication showed trend to shorten.
Mierzyńska et al. (Thu,) studied this question.