Risk 152

Course overview

Hour 1 *Analyze ethical and legal challenges where human healthcare professional judgment diverges from that of deep learning medical hardware, and employment of "off label" use of AI-assisted devices. *Evaluate "Black Box" and "hallucination" liability dilemmas in AI-assisted healthcare, outlining parties' accountability for AI error. *Explain the foundational risk of algorithmic bias, using the "Garbage In, Garbage Out" principle to describe its impact on AI effectiveness in healthcare. *Distinguish between fixed and continuously-learning algorithms, how the FDA treats them for purposes of healthcare device approval, and the effect on professionals' liability. *Contrast consumer data protection provided by the FTC with the patient-data protection by HIPAA and explain the "Privacy Gap" for consumer health devices. Hour 2 *Differentiate the roles of international bodies (like ISO) and U.S. national agencies (like NIST) in creating safe AI standards. *Evaluate legal precedents (such as the Washington teeth whitening case) and discuss why a professional is held to a higher legal standard of care than a layperson, even when performing the same task, and the importance of preventing "public confusion" regarding those standards. *Explain how the FDAs rigorous approval process for diagnostic devices serves to "supplant" or "substitute" for the individual judgment of a licensed healthcare professional. *Identify the three core elements that generally constitute the practice of medicine under U.S. state laws and discuss the public safety risks that occur when a patient receives an accurate diagnosis from a machine but lacks the counseling or context provided by a human professional. *Explain the economic and regulatory incentives for AI manufacturers to move operations to countries with less rigorous approval processes to provide "lifesaving" care to international consumers.