Projects per year
Abstract
Artificial intelligence (AI) has the potential to transform healthcare, but its clinical use also has important challenges and limitations. Recently natural language processing and generative pre-training transformer (GPT) models have gained particular interest due to their ability to simulate human conversation. We aimed to explore output of the ChatGPT model (OpenAI, https://openai.com/blog/chatgpt) regarding current debates in cardiovascular CT. Prompts included debate questions from the Society of Cardiovascular Computed Tomography 2023 programme as well as questions about high risk plaque (HRP), quantitative plaque analysis, and how AI will transform cardiovascular CT. The AI model rapidly provided plausible responses including both pro and con sides of the argument. Advantages of AI for cardiovascular CT that were described by the AI model included improving image quality, speed of reporting, accuracy, and consistency. The AI model also acknowledged the importance for continued involvement of clinicians in patient care.
Original language | English |
---|---|
Journal | Journal of Cardiovascular Computed Tomography |
Early online date | 7 Apr 2023 |
DOIs | |
Publication status | E-pub ahead of print - 7 Apr 2023 |
Fingerprint
Dive into the research topics of 'How will artificial intelligence transform cardiovascular computed tomography? A conversation with an AI model '. Together they form a unique fingerprint.Projects
- 1 Active
-
Incidental coronary calcification on thoracic computed tomography
Williams, M. (Principal Investigator), Mills, N. (Co-investigator) & Newby, D. (Co-investigator)
1/02/21 → 31/01/26
Project: Research