Foundation artificial intelligence models such as ChatGPT are machine models which have gained renown over the last few years but have yet to be utilized in cardiovascular medicine. These models have been trained by humans and are capable of performing a variety of tasks such as interpretation of text, images and dataset. Equally, they can assimilate a variety of different inputs from which they can learn and predict patterns in data. Furthermore, they have the ability to learn and transfer skills from one task to another.
Current AI models in medicine are rigid and designed for very specific, repetitive tasks and cannot be used outside of their original context. Despite these limitations, they have already proven useful in medicine. We propose to generate a new, powerful form of AI (“CardioSage”) which aims to take health information from multiple sources to help in the diagnosis and prevention of cardiovascular disease. The results of this are likely to be wide-ranging and may augment healthcare worker’s ability to detect and treat heart disease in a way that is impossible for humans to achieve on their own. We anticipate that this will lead to more accurate prediction of disease presence and progression and will empower healthcare delivery around the world and help to reduce the burden of cardiovascular disease around the world. We anticipate that the project will take 36 months to complete.