The applications of artificial intelligence (AI) in perioperative cardiothoracic and vascular practice continue to grow, given its multiple applications from discovery across translation and integration to delivery of clinical care.^1-3 These diverse applications have included risk prediction, decision support, perioperative imaging, educational delivery, and research trials, as well as academic publishing, to such an extent that they significantly affect daily workflow and clinical practice.^1-6 Although these applications offer important advances, challenges remain as AI is integrated into our specialty of cardiothoracic and vascular anesthesiology and critical care.^1,2 These challenges must be managed to deliver real-world effectiveness with high-performing models that can be easily implemented with current workflows.^1,2 Furthermore, these numerous applications of AI should be deployed in an equitable fashion across the clinical, educational, and research domains with robust regulatory oversight to accelerate innovation and dissemination.^1,2,6

The assisted and autonomous care possibilities offered by AI are also particularly attractive in our specialty given the current workforce crisis.^2,7 With the current imbalance across supply and demand in the cardiovascular workforce, AI can assist with closing this chasm as we navigate this challenging landscape after the pandemic era.^7-9 The deployment of suitable machine learning models could assist with the unloading of mundane tasks in the workforce arena.^8,9 Further development of educational solutions with deep learning techniques could augment the planning and delivery of care as well as the learning of important procedural skills.^1,8 The tailored implementation of AI also could enhance the ability to supervise delivery of care with greater efficiency and safety.⁸ Likewise, the supervised integration of solutions from the AI landscape could unload the current workforce to assist with recruitment and retention to bridge the current gaps in supply and sustainability.^1,2,6-9

With the ongoing crisis in the cardiovascular workforce, major gaps in the availability and quality of care have developed, with a widening gulf between evidence and practice.^9-11 In the cardiovascular arena, heart failure represents such a domain: it is a serious condition with a rising prevalence, a falling life expectancy, workforce shortages, and multiple gaps in care delivery.^10-13 Given this challenging landscape, the question has developed whether AI can deliver care in this high-priority domain in an autonomous fashion to address these failures in execution to bridge the gaps between therapeutic advances and delivery of care.¹⁰

The US government is currently funding the development of autonomous clinical care options for patients with heart failure through the Advanced Research Projects Agency for Health with the launch of the Agentic AI-Enabled Cardiovascular Care Transformation program.¹⁴ This initiative aims to foster the development of trustworthy clinical AI agents to adjust the medical management of patients with heart failure in areas such as appointments and medications, as well as lifestyle choices including diet and exercise.¹⁴ Furthermore, the program includes the development of a supervisory clinical AI agent to oversee these activities as a further check for safety and efficacy.¹⁴ This integrated cardiovascular care platform could provide real-time guidance for patients with heart failure and could be deployed in conjunction with health systems across urban and rural settings.