Artificial intelligence in cardiology: Exciting but handle with caution

Abstract

While artificial intelligence is not new, recent advances have caught the attention of people in many fields of study. In fact, Alan Turing, who conceived of the idea of what is now called a computer in 1937, was also an early leader in the field of artificial intelligence. The Turing test was developed to measure a machine's ability to be indistinguishable from human output [ [1] Topol E. Deep Medicine: How Artificial Intelligence Can Make Healthcare Human Again. Basic Books, 2015 Google Scholar ]. Only recently, with sufficient computing power, the internet, big data, and large language models, such as ChatGPT4 and has the vision of ‘AI’ begun to be met [ [2] Topol E.J. High-performance medicine: the convergence of human and artificial intelligence. Nat. Med. 2019; 25: 44-56 Crossref PubMed Scopus (1969) Google Scholar ]. In addition to ChatGPT, which as developed by OpenAI, in collaboration with Microsoft, other large language model, possessing large data and computing power have been developed by Google and Facebook, and it is likely that more will follow. ChatGPT4 recently gained a lot attention when, despite no formalized training, was able to pass the United States Medical Licensing examination (USMLE), providing explanations of the medical reasoning [ [3] Nori H. King N. McKinney S. Carignan D. Horvitz E. Capabilities of GPT-4 on medical challenge problems. arXiv. 20 Mar 2023; (2303.13375v1 [cs.CL]) Google Scholar ]. Artificial Intelligence and machine learning systems (AI/ML) are only as good as the dataset that was used to train the system. Bias in the training dataset will be passed onto the system.