Artificial Intelligence within Healthcare

Privacy PubMed Search

Additional Readings:

• Ethical and Legal Challenges of Artificial Intelligence-Driven Healthcare - Social Science Research Network
• AI Chatbots, Health Privacy, and Challenges to HIPAA Compliance - JAMA
• AI Productivity Tools and PHI - UA College of Medicine Tucson - ITS
• Approved (PHI) Solutions - UA College of Medicine Tucson - ITS

Ethics PubMed Search

Websites: 

• Federal AI Ethical Standards and legal Nuances - University of Arizona
• WHO calls for safe and ethical AI for Health - World Health Organization
• Advancing healthcare AI through ethics, evidence and equity - American Medical Association

Multimedia:

• Ethics, Bias & Artificial Intelligence in Healthcare - University of Colorado, Department of Medicine
• Medical ethics and evidence-based medicine - American Medical Association
• Safe, Fair and Ethical AI in Healthcare - Standford Center for Artificial Intelligence in Medicine & Imaging

Bias PubMed Search

Artificial intelligence requires input of data in order to advance its training. Some issues concern where the data may come from, ensuring patient privacy, and how to standardize regulations across all models.

Websites and Multimedia:

• Shedding Light on Healthcare Algorithmic and Artificial Intelligence Bias - US Department of Human and Health Services: Office of Minority Health
• Considerations for addressing bias in artificial intelligence for health equity - PubMed
• The Promise of Artificial Intelligence and Machine Learning for Improving Clinical Outcomes - Digital Medicine Society (DiMe)

• Advancing Health Equity with Medical Technologies - FDA

Currently, the FDA has established a framework for evaluating AI-based medical devices focusing on premarket review, post-market surveillance, and continuous learning. 

AI models are assessed based on their intended use, risk to patients, and the nature of their algorithm. Diagnostic AI tools, which directly inform clinical decision without human intervention, carry a higher risk if they fail or provide inaccurate results. Consequently, they require comprehensive validation and monitoring to ensure their safety and effectiveness. 

The WHO guidance on Ethics and Governance of Artificial Intelligence for Health, developed over eighteen months by experts in various fields, addresses the promise of AI in improving health related functions such as diagnosis, treatment, and public health response. The report emphasizes that AI technologies must prioritize ethics and human rights, ethical challenges, and six principles to ensure AI benefits all countries, and provides recommendations for governing AI. 

Additional Resources:

• U.S. Artificial Intelligence Safety Institute Guidance - NIST