The AI Revolution in Medical Education
Journal of Education Practices and Trends, April 2025
Editor in Chief, Innovation and Entrepreneurship in Health
An excerpt of a white paper, this article explores key AI-driven innovations in medical education, addressing solutions for medical students, educators, and administration.
Table of Contents
I. Abstract
II. The Revolution of AI
III. Current AI Applications in Medical Education
a. Solutions for Medical Students
i. AI-Powered
Personalized Learning
ii. Virtual Patients
and Simulations
iii. AI-Assisted
Diagnostic Training
iv. Natural Language
Processing for Medical Literature Review
b. Solutions for Medical Educators
i. AI for
Curriculum Design and Optimization
ii. Automated
Assessment and Feedback Systems
iii. AI-Enhanced
Lecture Preparation and Delivery
iv. Virtual Teaching
Assistants
c. Solutions for Administration
i. Predictive
Analytics for Student Performance and Support
ii. AI-Driven
Admissions Processes
iii. Resource
Allocation and Scheduling Optimization
iv. Research Collaboration
and Funding Opportunity Identification
IV. Current AI Applications in Medical Education
a. AI in Clinical Skills
Assessment
b. Augmented and Virtual Reality
in Medical Training
c. AI-Powered Adaptive Testing
for Medical Licensing Exams
d. Continuous Learning and Competency
Assessment for Practicing
Physicians
V. Challenges and Considerations
a. Ethical Considerations in AI
Use for Medical Education
b. Data Privacy and Security
Concerns
c. Balancing AI Integration with Human
Expertise and Empathy in Medical
Training
d. Addressing These Challenges
Requires a Multifaceted Approach
e. Infrastructure and Technology
Requirements
VI. Preparing for the AI-Driven Future of Medical
Education
a. Faculty Development and AI
Literacy
b. Curriculum Redesign to
Incorporate AI Competencies
c. Collaboration Between Medical
Schools, Tech Companies, and
Healthcare Providers
d. Implementing These Preparatory
Measures Requires a Strategic
Approach
VII. The Role of Public-Private Partnerships (PPPs) in
Advancing AI in Medical
Education
a. Models for Public-Private
Collaboration
b. Implementing Effective
PPPs
c. Challenges and
Considerations
VIII. Conclusion
IX. References
Abstract
The artificial intelligence (AI) revolution is transforming medical education at an unprecedented pace, offering game-changing opportunities to personalize learning experiences, support medical educators in their daily tasks, and optimize education management in medical schools and teaching hospitals. Recent research demonstrates the immense potential of AI to boost productivity in medical knowledge work, with studies showing that medical professionals using advanced AI models completed more tasks, worked faster, and produced higher-quality outputs compared to those without AI assistance. This article explores key AI-driven innovations in medical education, addressing solutions for medical students, educators, and administration.
The Revolution of AI
AI is rapidly transforming healthcare and medical education, reshaping the training of future healthcare professionals. As these technologies grow more sophisticated and accessible, their potential to personalize medical curricula and simulate complex patient scenarios becomes increasingly clear, promising to enhance educational outcomes and empower medical educators in preparing students for an evolving healthcare landscape. Globally, leading medical institutions are adopting unique approaches to leverage AI in revolutionizing medical education according to their specific goals and healthcare challenges. For instance, Singapore’s Smart Nation strategy aims to establish the country as a world leader in AI by 2030, incorporating extensive AI use to support personalized education and improve care for patients with complex medical conditions. South Korea is also personalizing training simulations and learning activities to align with individual skill levels and learning styles. At Stanford University (Stanford Medicine, 2024), the AI-Med program integrates AI throughout its medical curriculum, focusing on both practical applications and the ethical implications of AI in healthcare. Meanwhile, the University of Florida has partnered with NVIDIA to develop an AI-powered curriculum that includes training in applications like diagnostic imaging, personalized medicine, and healthcare robotics.
As AI becomes increasingly prevalent in medical education, it is essential for policymakers, educational institutions, and healthcare stakeholders to recognize both its potential benefits and the challenges associated with its implementation. Dr. Eric Topol (2022) posits that AI may represent the most significant shift that has ever occurred in medicine and medical education. Yet, concerns exist about how AI integration might affect the cultivation of essential human skills, such as empathy and clinical reasoning. The rising use of AI-powered tools like ChatGPT in medical training raises issues of over-reliance on technology, academic integrity, and the necessity for AI-detection tools, alongside the development of robust ethical frameworks. A recent survey by the Association of American Medical Colleges (2024) indicates that while medical schools are adapting to the increasing presence of AI, significant concerns remain about preserving the human element in education and ensuring that AI tools do not reinforce existing gaps in healthcare delivery.
This article examines the global state of AI in medical education, focusing on applications in medical schools and teaching hospitals. It is crucial to emphasize that AI should not be viewed as a replacement for human expertise in training, but rather as a means to enhance and amplify human judgment and clinical skills. The article will identify current AI applications that improve clinical skills training, facilitate medical knowledge acquisition, assist in diagnostic reasoning, and offer personalized learning experiences. Additionally, it will discuss the role of AI in educational management and the importance of public-private partnerships in fostering innovation.
Current AI Applications in Medical Education
AI’s rapid integration into medical education has led to innovative applications designed to enhance teaching, learning, and education management. Key AI-driven advances provide solutions in medical schools and teaching hospitals for medical students, educators, and administrators.
Solutions for Medical Students
AI-Powered Personalized Learning
AI-driven adaptive learning platforms are transforming how medical students acquire and retain knowledge. These AI-powered tutoring systems customize instruction to individual needs, offering real-time feedback across various medical disciplines. By using machine learning algorithms to analyze performance and identify knowledge gaps, they create personalized learning paths. AMBOSS, a comprehensive medical learning platform, adapts its question bank and resources based on each student’s performance data, tailoring content delivery to target areas needing improvement.
Virtual Patients and Simulations
AI-powered simulations and virtual patients provide medical students with realistic training environments to practice clinical skills and decision-making safely. These immersive experiences enable students to interact with virtual patients, gather medical histories, conduct examinations, and develop treatment plans while receiving immediate AI feedback. Oxford Medical Simulation (OMS) offers VR-based clinical simulations where AI adapts patient responses based on student actions, creating a dynamic learning experience.
AI-Assisted Diagnostic Training
AI-powered diagnostic tools are being integrated into medical education to enhance students’ clinical reasoning skills. These tools analyze medical images, lab results, and patient data to generate differential diagnoses, suggest further investigations, and provide relevant educational resources. By using these tools, students can learn from real-world scenarios and improve their diagnostic accuracy. Enlitic, an AI company focused on medical imaging, has developed tools for radiology education that highlight areas of interest in images and explain the reasoning behind diagnoses, helping students refine their interpretative skills.
Natural Language Processing for Medical Literature Review
The vast amount of medical literature can be overwhelming for students. Natural language processing (NLP) techniques are being used to develop AI-powered tools that efficiently search, summarize, and extract relevant information from medical literature. These tools help students stay updated on the latest research, identify relevant clinical trials, and access evidence-based information for their learning and clinical decision-making. IBM’s Watson for Oncology (MIT-IBM Watson AI Lab, (2024) is utilized in some medical schools to teach students how to quickly synthesize information from thousands of medical journals and clinical trials when researching treatment options for cancer patients.
Solutions for Medical Educators
AI for Curriculum Design and Optimization
AI analytics are being used to analyze student performance data and feedback to continuously improve and optimize medical curricula. The University of Michigan Medical School is using AI to analyze student performance data across various courses and clinical rotations to identify areas where the curriculum can be improved or better integrated.
Automated Assessment and Feedback Systems
AI-powered systems are increasingly utilized to deliver more frequent and objective assessments of student performance in clinical reasoning and decision-making. These systems analyze teaching sessions, offering educators valuable feedback on their communication styles and overall effectiveness, which helps them identify areas for improvement. Shadow Health’s Digital Clinical Experiences leverage AI to evaluate students’ interactions with virtual patients, providing detailed insights into their communication skills, clinical reasoning, and empathy.
AI-Enhanced Lecture Preparation and Delivery
AI tools are assisting medical educators in creating more engaging and effective lectures by suggesting relevant content, generating visual aids, and even providing real time analytics on student engagement during lectures. Presentation AI, while not specific to medical education, is being adopted by some medical educators to enhance their lecture slides with AI-generated content and visuals.
Virtual Teaching Assistants
AI-powered chatbots and virtual assistants are being used to provide 24/7 support to students by answering common questions and guiding course material. This enables educators to focus more on complex tasks such as curriculum development and personalized mentoring. Georgia Tech’s Jill Watson, initially developed for computer science, is being adapted for medical education to address routine student inquiries and explain basic medical concepts.
Solutions for Administrators
Predictive Analytics for Student Performance and Support
AI algorithms are being used to identify students who may be at risk of academic struggles or burnout, allowing for early intervention and support. The University of Arizona College of Medicine is using an AI-driven early warning system to identify students who may be at risk of failing key exams or experiencing mental health issues.
AI-Driven Admissions Processes
Some medical schools are experimenting with AI to assist in the admissions process by helping screen applications. While not yet widely implemented, companies like Kira Talent are developing AI-enhanced admissions tools that could be adapted for medical school admissions to help assess soft skills and promote consistent evaluations.
Resource Allocation and Scheduling Optimization
AI is being used to optimize the complex scheduling of clinical rotations, lectures, and use of simulation facilities in medical schools. The University of Texas Health San Antonio is using AI-powered scheduling software to optimize the allocation of clinical rotation slots, ensuring students get a diverse range of experiences while balancing the needs of various hospital departments.
Research Collaboration and Funding Opportunity Identification
AI tools are being employed to identify research collaboration opportunities and match researchers with potential funding sources. Dimensions, an AI-powered research insights platform, is being used by some medical schools to help faculty identify funding opportunities and potential collaborators for medical research projects. These AI applications are transforming medical education, offering personalized learning experiences, enhancing clinical skills training, and optimizing administrative processes. As these technologies continue to evolve, they have the potential to significantly improve the quality and efficiency of medical education, ultimately leading to better-prepared healthcare professionals and improved patient care.
Emerging Trends and Future Applications
As AI advances, new applications in medical education are transforming how healthcare professionals are trained and assessed, with key trends shaping the future of medical education.
AI in Clinical Skills Assessment
AI-powered systems are being developed to provide more objective and comprehensive assessments of clinical skills. These systems use computer vision, natural language processing, and machine learning to analyze student performance in simulated clinical scenarios. can assess medical students’ performance during Objective Structured Clinical Examinations (OSCEs). The system uses video analysis to evaluate students’ body language, facial expressions, and physical examination techniques, while speech recognition and NLP analyze their communication with standardized patients. As these systems become more sophisticated, they could provide more consistent and detailed feedback on clinical skills, potentially reducing variability in assessments and allowing for more frequent, low-stakes evaluations throughout medical training.
Augmented and Virtual Reality in Medical Training
While VR and AR are currently utilized in medical education, future applications are expected to feature more advanced AI integration, resulting in highly realistic and adaptive virtual training environments. Microsoft’s HoloLens, paired with AI, is already creating mixed reality experiences for anatomy education, with future versions potentially adjusting anatomical model complexity based on students’ understanding and learning styles. AI-enhanced AR and VR could facilitate immersive training in rare or high-risk scenarios, deliver real-time information during procedures, and provide personalized feedback in a risk-free environment, significantly improving procedural skills training and patient safety.
AI-Powered Adaptive Testing for Medical Licensing Exams
AI is poised to transform medical licensing exams by enabling more adaptive and personalized assessments. These systems adjust question difficulty based on an examinee’s performance, providing a more accurate evaluation of their knowledge and skills. The National Board of Medical Examiners (NBME) is investigating AI’s potential for developing adaptive testing for the United States Medical Licensing Examination (USMLE), utilizing algorithms to modify questions in real time. This approach could enhance assessment accuracy, reduce test anxiety, improve exam validity, and facilitate more frequent, shorter evaluations throughout a physician’s career to ensure ongoing competency.
Continuous Learning and Competency Assessment for Practicing Physicians
AI systems are being developed to support lifelong learning and continuous assessment of practicing physicians, helping them stay current with medical knowledge and maintain clinical skills. The American Board of Internal Medicine is piloting a program called Knowledge Check-In, which uses AI to offer personalized learning recommendations based on physicians’ performance in regular assessments. In the future, AI could analyze a physician’s clinical decisions and patient outcomes to provide real-time feedback and targeted learning suggestions, identifying areas for additional training or support and ensuring high-quality patient care.
As these trends evolve, they have the potential to create a more personalized, efficient, and effective medical education system throughout a physician’s career. However, it is crucial to consider the ethical implications, ensure the validity and reliability of these technologies, and balance AI assessments with human judgment in evaluating medical competence. The successful integration of these AI applications will require ongoing collaboration between medical educators, AI researchers, and regulatory bodies to enhance, rather than replace, the essential human elements of medical training and practice. While AI has great potential to improve medical education, its integration poses significant challenges that need careful attention.
Challenges and Considerations
Ethical Considerations in AI Use for Medical Education
The use of AI in medical education raises important ethical questions that educators and institutions should consider, including how to ensure that AI-driven decisions in education are explainable and transparent, who is responsible when AI makes errors in assessment or recommendation, and how to ensure students understand and consent to AI’s role in their education. In response to these concerns, Stanford University’s AI Ethics in Medical Education Initiative (Stanford Medicine, 2024) is developing guidelines for the ethical use of AI in medical training, addressing issues such as algorithmic predisposition, AI hallucination, and the preservation of human judgment in clinical decision-making.
Data Privacy and Security Concerns
The use of AI in medical education involves handling sensitive student data and potentially patient information, which raises critical privacy and security concerns, such as how to safeguard student and patient data used in AI systems, who owns the data generated by AI educational tools, and how to manage data in global educational collaborations. In response to these challenges, the University of California San Francisco (UCSF) has implemented strict data governance policies for its AI initiatives in medical education. These policies include the anonymization of student data and secure, audited access controls to ensure the protection of sensitive information.
Balancing AI Integration with Human Expertise and Empathy in Medical Training
While AI can enhance many aspects of medical education, it is crucial to maintain a focus on developing essential human skills, raising important questions such as how to ensure AI does not diminish the emphasis on empathy and interpersonal skills, how AI can support rather than replace the development of clinical judgment, and how to prevent students from becoming overly dependent on AI tools. A prime example of this balanced integration is the Mayo Clinic Medical School’s introduction of a Human Skills in the Age of AI curriculum, which emphasizes developing empathy, communication, and ethical decision-making skills alongside AI literacy.
Addressing These Challenges Requires a Multifaceted Approach
By proactively focusing on the challenges of AI integration, medical education can harness the benefits of AI while upholding the core values and essential human elements of the medical profession. Medical schools should establish clear guidelines for AI use in education, addressing issues of transparency, accountability, and consent. Institutions must adopt stringent data protection measures and clear policies on data ownership and usage. Collaboration with diverse stakeholders must occur to ensure AI tools are developed with consideration for all student populations. Design curricula must balance AI integration with strong focus on empathy, communication, and clinical reasoning skills. Partnerships between medical educators, ethicists, AI researchers, and policymakers must be fostered to address these complex challenges. Continuous evaluation and adaptation are essential, regularly assessing the impact of AI tools on educational outcomes and student well-being, adapting approaches as needed. This balanced approach will be crucial in preparing the next generation of healthcare professionals to thrive in an AI-enhanced medical landscape while maintaining the human touch that is central to quality patient care.
Infrastructure and Technology Requirements
To effectively integrate AI into medical education, institutions need robust technological infrastructure. High-speed internet connectivity is necessary to ensure reliable, high-bandwidth connections to support AI-powered applications and simulations. Investments in advanced, powerful computing systems capable of running complex AI models and processing large datasets must be made. Protective data storage should be implemented to secure scalable data storage solutions and manage the vast amounts of data generated by AI educational tools. State-of-the-art simulation centers equipped with AI-enhanced virtual and augmented reality technologies need to be developed. Johns Hopkins University School of Medicine has established an AI Innovation Lab, equipped with high-performance computing clusters and advanced visualization tools, dedicated to developing and testing AI applications for medical education and research.
Preparing for the AI-Driven Future of Medical Education
Faculty Development and AI Literacy
Preparing faculty for the AI revolution is crucial for successful integration. AI literacy programs are needed to develop comprehensive training programs to ensure faculty understand AI fundamentals, applications, and limitations. Hands-on experience provides opportunities for faculty to work with AI tools in their teaching and research. Ethical AI use will train faculty on the ethical considerations and best practices for using AI in medical education. In addition, continuous learning establishes ongoing professional development programs to keep faculty updated on the latest AI advancements. Harvard Medical School has launched an AI in Medicine certificate program for faculty, covering topics from basic AI concepts to advanced applications in medical education and clinical practice.
Curriculum Redesign to Incorporate AI Competencies
Medical curricula must evolve to prepare students for an AI-enhanced healthcare landscape. One factor is integrating basic AI concepts and applications into core medical curricula. Data science skills must also be incorporated by training in data analysis, statistics, and programming to enable students to work effectively with AI tools. AI ethics and governance should be included in courses, including the ethical implications of AI in healthcare and the importance of responsible AI use. AI-augmented clinical skills must also be developed via new modules that teach students how to use AI tools in diagnosis, treatment planning, and patient care. An illustration of curriculum redesign is Stanford University School of Medicine’s introduction of a Medical AI and Computer Vision track within its M.D. program, allowing students to develop expertise in AI applications alongside traditional medical training.
Collaboration Between Medical Schools, Tech Companies, and Healthcare Providers
Fostering partnerships is essential for developing relevant, cutting-edge AI solutions. Industry partnerships allow collaboration with tech companies to develop and test AI tools tailored for medical education. Healthcare system integration should work closely with healthcare providers to ensure AI education aligns with real-world clinical needs. Research collaborations can establish joint research initiatives between medical schools and AI research centers to advance the field of AI in healthcare education. Shared resources can create consortiums to share costs, data, and expertise in AI development and implementation. For example, the University of California San Francisco (UCSF) has partnered with Google Health to develop AI tools for medical imaging education, combining UCSF’s clinical expertise with Google’s AI capabilities.
Implementing These Preparatory Measures Requires a Strategic Approach
By taking certain steps, medical education institutions can position themselves at the forefront of the AI revolution in healthcare, ensuring that future healthcare professionals are not only comfortable with AI technologies but also capable of leading their development and ethical application in clinical practice. This proactive approach involves conducting a comprehensive AI readiness assessment to identify current capabilities and gaps, developing a long-term AI integration roadmap with clear milestones and resource allocation plans, and establishing an AI governance committee to oversee implementation while ensuring alignment with institutional goals and ethical standards. Additionally, creating feedback mechanisms will allow for continuous evaluation and refinement of AI integration efforts, and fostering a culture of innovation and adaptability among faculty and students will further enhance this initiative. The successful integration of AI into medical education will ultimately lead to better-prepared healthcare professionals, more efficient healthcare systems, and improved patient outcomes. However, it’s crucial to remember that AI should augment, not replace, the human elements of medical practice. The goal is to create a symbiotic relationship between AI and human expertise, leveraging the strengths of both to advance medical education and healthcare delivery.
The Role of Public-Private Partnerships (PPPs) in Advancing AI In Medical Education
Fostering strong public-private partnerships (PPPs) is crucial for accelerating the development and adoption of AI technologies in medical education. The complexity and resource-intensive nature of AI innovation in healthcare necessitates pooling expertise, resources, and investment from both the public and private sectors. PPPs leverage complementary expertise by combining the strengths of each partner: the public sector’s understanding of healthcare policies and educational standards, the private sector’s technological capabilities and agility, and academia’s research expertise. This collaboration optimizes resources, as shared investments reduce financial burdens on individual institutions and enable the development of more sophisticated AI solutions. Additionally, it ensures real-world relevance by aligning AI tools with both clinical needs and educational objectives. Furthermore, PPPs accelerate innovation by streamlining the path from research to implementation, allowing cutting-edge AI tools to be integrated into medical education more rapidly.
Models for Public-Private Collaboration
Joint research and development initiatives, such as the MIT-IBM Watson AI Lab (2024), exemplify collaboration in creating AI tools for medical education by combining IBM’s AI expertise with clinical insights from academic medical centers. Data sharing and open standards are illustrated by the UK Biobank project, a partnership between the NHS, universities, and private companies that provides anonymized health data for AI research in both medical education and clinical applications. Innovation hubs, like the Houston Methodist-Amazon Web Services (AWS) Innovation Hub, focus on developing AI applications for medical training and clinical decision support. Additionally, AI education consortiums, such as the AI-READI (Artificial Intelligence Ready and Equitable Atlas for Diabetes Insights), funded by the NIH, bring together universities, tech companies, and community organizations to create diverse datasets for AI training in diabetes care. Lastly, clinical integration partnerships, such as Google Health’s collaboration with Mayo Clinic, aim to incorporate AI tools into medical student training and residency programs.
Implementing Effective PPPs
To ensure the success of public-private partnerships (PPPs) in medical education, it is essential to establish clear governance structures that promote transparent decision-making processes and define roles and responsibilities for all partners. Additionally, developing comprehensive ethical frameworks is crucial to address issues, such as data privacy and responsible AI use. Intellectual property agreements should be created to outline policies on ownership and licensing of jointly developed AI tools and technologies. Engaging the community by involving patient advocacy groups and community representatives helps ensure that AI solutions meet diverse healthcare needs. Finally, implementing continuous evaluation through regular assessments of PPP outcomes will allow for adjustments in strategies to maximize their impact on medical education.
Challenges and Considerations
Balancing profit motives with educational goals is crucial to ensure that commercial interests do not overshadow the primary objective of enhancing medical education. It is equally important to implement robust safeguards for sensitive medical data used in AI development and education, ensuring data privacy and security. Strategies must be developed to promote equity in access, ensuring that AI tools are available to a wide range of medical institutions, not just those that are well-funded. Additionally, navigating complex healthcare regulations while fostering innovation in AI-driven medical education is essential for creating a sustainable and effective educational environment.
Conclusion
The AI revolution in medical education presents unique opportunities and complex challenges for the global healthcare community. AI technologies are poised to transform every aspect of training future healthcare professionals, including personalized learning experiences, advanced clinical simulations, and data-driven curriculum design. This transformation has the potential to significantly enhance the effectiveness and efficiency of medical education, which could lead to better-prepared healthcare professionals and improved patient outcomes.
The multifaceted applications of AI, ranging from AI-powered tutoring systems to virtual patient simulations and automated assessment tools, are diverse and rapidly evolving. As AI is integrated into medical training, ethical considerations are to be prioritized, including data privacy, algorithmic fairness, and the preservation of human empathy in healthcare. Furthermore, medical institutions must invest in robust technological infrastructure and faculty development to fully harness the benefits of AI in education. Evolving medical curricula to incorporate AI competencies is essential to ensure that future physicians are not only proficient users of AI but also informed shapers of its development in healthcare. A collaborative approach through public-private partnerships will be crucial in driving innovation and ensuring that AI tools are relevant, effective, and accessible to all medical education experiences.
The future of medical education will likely be characterized by a symbiotic relationship between AI technologies and human expertise. AI will augment and enhance the learning process, allowing for more personalized, efficient, and comprehensive medical training. However, it will be crucial to maintain a balance, ensuring that the essential human elements of medical practice – empathy, ethical reasoning, and clinical judgment – remain at the forefront of medical education. Therefore, a call to action is required to successful achieve such a balance.
Medical educators must embrace AI as a powerful tool to enhance one’s teaching but remain vigilant in fostering the human skills that define great physicians. Medical institutions must invest in AI infrastructure and faculty development, while also establishing robust ethical frameworks for AI use in education. Policymakers need to develop regulations that encourage innovation in AI-driven medical education while safeguarding student and patient interests. Technology developers should collaborate closely with medical educators to create AI tools that truly address the needs of medical training and align with healthcare values. Students must engage actively with AI technologies not only as part of their learning but also to cultivate the uniquely human skills that sets them apart as a healthcare professional.
The AI revolution in medical education is not a distant future – it is unfolding now. By proactively addressing challenges, fostering collaboration, and maintaining a focus on ethical and human-centered design, one can harness the power of AI to create a new paradigm in medical education. This paradigm will not only produce more knowledgeable and skilled healthcare professionals but also cultivate compassionate, ethical, and adaptable physicians ready to lead in an AI-enhanced healthcare landscape. The journey ahead is complex, but the potential rewards – in terms of improved medical education, enhanced patient care, and accelerated medical progress – are immense. As AI-driven transformation occurs, let everyone uphold core values of medicine while embracing its potential to shape a brighter future for healthcare education and practice.
References
https://www.ama-assn.org/practice-management/digital/augmented-intelligence-medicine
https://www.aamc.org/what-we-do/mission-areas/medical-education/ai-in-medical-education
https://mitibmwatsonailab.mit.edu/category/healthcare-and-life-sciences/
https://www.nhsx.nhs.uk/ai-lab/
https://med.stanford.edu/ai.html
http://www.royalcollege.ca/rcsite/documents/ai-task-force/ai-task-force-e.pdf
https://www.who.int/publications/i/item/9789240029200