Clinical Report: AI’s Expanding Role in Resuscitation

Overview

The article describes AI tools for CPR education and emergency guidance, focusing on ChatCPR, which scored higher than 911 dispatchers on guideline-based CPR checklists in a retrospective proof-of-concept study. It also discusses technology-enabled manikin training and barriers that may deter bystanders from performing CPR. The authors state that further testing is needed before ChatCPR is used in real-world emergencies.

Background

The source reports that approximately 90% of people experiencing out-of-hospital cardiac arrest do not survive and that immediate CPR can triple an individual’s chance of survival. Bystander CPR increased from 41.7% in 2024 to 42.5% in 2025, but some bystanders hesitate because they fear lacking skills or exposing a female patient’s breasts. Thomas Sather of the American Red Cross Scientific Advisory Council emphasized that emergency performance depends on training and repeated practice, as skills can deteriorate without it. The article presents AI-guided instruction and feedback-enabled training as approaches being explored to support CPR performance and learning.

Data Highlights

Key Findings

• The article reports that immediate CPR can triple an individual’s chance of survival after out-of-hospital cardiac arrest, while approximately 90% of people experiencing such an arrest do not survive.
• Some bystanders report fear of inadequate skills and concerns about exposing a female patient’s breasts as reasons for hesitating to provide CPR.
• In the reported assessment, publicly available AI models averaged approximately 90% for essential CPR instructions and approximately 70% for advanced instructions; the team used contextual prompt engineering and dispatcher-manual guidance to develop ChatCPR.
• ChatCPR achieved 100% on guideline-based CPR checklists. In a retrospective comparison, it met approximately 100% of basic and 99% of advanced recommendations, compared with approximately 85% and 63%, respectively, for dispatchers.
• The researchers do not propose replacing emergency dispatchers; they describe AI as a possible aid to dispatchers or an option when a dispatcher cannot be reached immediately, and state that further testing is required before real-world use.
• The article describes the Elevate SMART Manikin as providing objective feedback on first aid, CPR, automated external defibrillator, and basic life support skills using evidence-based algorithms. A related randomized trial report concerns AI- and sensor-integrated versus face-to-face CPR retraining, but the supplied context does not provide its outcome data.

Clinical Implications

The article describes ChatCPR as a potential adjunct to dispatcher-assisted CPR, not a replacement, and notes that further testing is required before emergency deployment. It also describes feedback-enabled manikins as a training tool; the supplied material does not establish their effect on patient outcomes.

Conclusion

The source presents AI-guided CPR and technology-supported training as developing approaches to resuscitation education and bystander assistance. ChatCPR’s reported checklist performance is promising within a retrospective proof-of-concept study, but the authors call for further testing before real-world emergency use.

Related Resources & Content

1. DIGITAL HEALTH, 2025 — Immediate learning outcomes of AI- and sensor-integrated versus face-to-face CPR retraining for chest compression quality: A randomized trial
2. American Heart Association, 2025 — Part 4: Systems of Care: 2025 American Heart Association Guidelines for Cardiopulmonary Resuscitation and Emergency Cardiovascular Care
3. American Heart Association, 2025 — Highlights of the 2025 American Heart Association's Guidelines for CPR and ECC
4. International Liaison Committee on Resuscitation, 2024 — 2024 International Consensus on Cardiopulmonary Resuscitation and Emergency Cardiovascular Care Science With Treatment Recommendations
5. DIGITAL HEALTH — Reflections on AI- and sensor-integrated versus face-to-face CPR reskilling for chest compression eminence
6. JAMA Network Open — Augmented Reality and the Shared Mental Model in Pediatric Resuscitation
7. DIGITAL HEALTH — Leveraging Artificial Intelligence in Cardiac Arrest Management: Transitioning from Algorithmic Efficacy to Integrated Digital Health Solutions Focused on Patient Outcomes
8. Part 4: Systems of Care: 2025 American Heart Association Guidelines for Cardiopulmonary Resuscitation and Emergency Cardiovascular Care
9. Highlights of the 2025 American Heart Association's Guidelines for CPR and ECC
10. 2024 International Consensus on Cardiopulmonary Resuscitation and Emergency Cardiovascular Care Science With Treatment Recommendations
11. Immediate learning outcomes of AI- and sensor-integrated versus face-to-face CPR retraining for chest compression quality: A randomized trial
12. Treatment of Cardiac Arrest | American Heart Association
13. Effect of Smartphone Dispatch of Volunteer Responders on Automated External Defibrillators and Out-of-Hospital Cardiac Arrests: The SAMBA Randomized Clinical Trial | Trials | JAMA Cardiology | JAMA Network
14. Smartphone‐activated volunteer responders and survival to discharge after out‐of‐hospital cardiac arrests in Victoria, 2018–23: an observational cohort study - PMC
15. 2025 ILCOR Statement
16. Effect of Machine Learning on Dispatcher Recognition of Out-of-Hospital Cardiac Arrest During Calls to Emergency Medical Services: A Randomized Clinical Trial | Emergency Medicine | JAMA Network Open | JAMA Network
17. Real-Time Smartphone Guidance Improves Cardiopulmonary Resuscitation (CPR) Performance in Trained and Untrained Individuals: A Stratified Simulation Study - PMC
18. Artificial intelligence in resuscitation: a scoping review - ScienceDirect
19. Generative artificial intelligence as a source of advice on resuscitation and first aid for laypeople: A scoping review - ScienceDirect