Artificial Intelligence (AI) is rapidly emerging as a pivotal tool in medical science, with its influence becoming increasingly prominent across a multitude of disciplines including the field of resuscitation.1,2 The year 2023 has been particularly notable for the expansion of AI applications, especially in the field of cardiac arrest (CA).3 This trend is reflected in the substantial growth of AI-related publications. We performed the following comprehensive analysis of literature on AI and CA to elucidate AI's actual role and its ongoing evolution. We searched and screened all relevant literature available on PubMed, with our search strategy detailed in the Supplemental Materials. After a manual review by two authors, ensuring a focused and relevant dataset, we selected 158 articles from a total of 6,109 identified on PubMed. We used ChatGPT-4 and Python as comparative controls to analyze all the data from PubMed. To categorize the content of these articles, we utilized Latent Dirichlet Allocation (LDA), an effective technique for identifying latent topics within textual data.4 The LDA analysis revealed
Trends and insights about cardiac arrest and artificial intelligence on PubMed using ChatGPT-4 / Semeraro, F.; Montomoli, J.; Cascella, M.; Bellini, V.; Bignami, E. G.. - In: RESUSCITATION. - ISSN 0300-9572. - 196:(2024). [10.1016/j.resuscitation.2024.110131]
Trends and insights about cardiac arrest and artificial intelligence on PubMed using ChatGPT-4
Semeraro F.;Bellini V.;Bignami E. G.
2024-01-01
Abstract
Artificial Intelligence (AI) is rapidly emerging as a pivotal tool in medical science, with its influence becoming increasingly prominent across a multitude of disciplines including the field of resuscitation.1,2 The year 2023 has been particularly notable for the expansion of AI applications, especially in the field of cardiac arrest (CA).3 This trend is reflected in the substantial growth of AI-related publications. We performed the following comprehensive analysis of literature on AI and CA to elucidate AI's actual role and its ongoing evolution. We searched and screened all relevant literature available on PubMed, with our search strategy detailed in the Supplemental Materials. After a manual review by two authors, ensuring a focused and relevant dataset, we selected 158 articles from a total of 6,109 identified on PubMed. We used ChatGPT-4 and Python as comparative controls to analyze all the data from PubMed. To categorize the content of these articles, we utilized Latent Dirichlet Allocation (LDA), an effective technique for identifying latent topics within textual data.4 The LDA analysis revealedI documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
