AI Medical Decisions: Where Liability Actually Sits

End-of-lesson check

15 questions · take it digitally for instant feedback at tendril.neural-forge.io/learn/quiz/end-ethics-safety-AI-medical-decision-accountability-adults

1. A hospital implements an AI diagnostic tool that the FDA has cleared as a Class II medical device. If the tool provides an incorrect diagnosis that leads to patient harm, where does primary legal liability rest?

   1. With the hospital as the deploying entity
   2. With the FDA, because the device received regulatory clearance
   3. With the AI vendor whose algorithm produced the error
   4. With the physician who relied on the AI output without independent verification

2. Which of the following represents the strongest informed consent practice regarding AI involvement in a patient's treatment?

   1. Documenting specific AI tool limitations and the physician's role in interpreting AI outputs
   2. Informing the patient that AI will be used without specifying the nature of its role
   3. Including a general statement in consent forms that 'advanced technology' may be used
   4. Requiring patients to sign a separate AI-specific consent waiver

3. A medical practice enters a vendor agreement for an AI triage system. The vendor contract includes a clause stating the vendor bears all liability for any adverse outcomes. In practice, who retains ultimate accountability for patient outcomes?

   1. The vendor, because the contract explicitly allocates liability
   2. The hospital administration, as the entity that purchased the system
   3. No one, because the contract creates a liability shield
   4. The physician, because professional accountability cannot be contractually transferred

4. Following an adverse event involving an AI-assisted diagnosis, what documentation element provides the most critical defense in potential malpractice litigation?

   1. The vendor's technical specifications for the AI tool
   2. Contemporaneous documentation of the physician's independent reasoning process
   3. The AI system's confidence score at the time of decision
   4. The patient's initial complaint and medical history

5. An AI tool used in a hospital has not received FDA clearance and is being used for research purposes only. What are the liability implications compared to an FDA-cleared tool?

   1. The physician faces increased liability exposure due to using an unapproved device
   2. The hospital assumes all liability, protecting the physician
   3. No difference in liability framework applies
   4. Liability is eliminated because research use is exempt from standard of care

6. When designing incident investigation processes for AI-involved adverse events, which element is essential for proper liability analysis?

   1. Relying solely on vendor-provided incident reports
   2. Documenting only the final clinical outcome, not the AI contribution
   3. Including systematic AI factor analysis alongside human decision review
   4. Removing AI factors from the investigation to focus on human error

7. A physician's current malpractice policy does not explicitly address AI-assisted procedures. What gap should be addressed with the insurance carrier?

   1. No action is needed since AI liability falls on vendors
   2. There may be coverage gaps for AI-specific claims or defense costs
   3. The physician should switch to an AI-specific policy that replaces malpractice coverage
   4. The policy likely already covers AI as standard medical practice

8. What is the fundamental principle regarding AI's role in medical decision-making liability?

   1. AI reduces malpractice exposure proportionally to its accuracy
   2. AI assists but cannot replace physician accountability
   3. AI eliminates the need for informed consent about technological tools
   4. AI can substitute for physician judgment in routine diagnostic cases

9. Patient-facing materials for an AI diagnostic tool should include which element to support both patient understanding and legal protection?

   1. Promotional claims about the tool's accuracy rates
   2. Comparison with other AI tools in the market
   3. Technical specifications of the AI algorithm's architecture
   4. Clear explanation of the tool's limitations and the physician's role in decisions

10. A hospital implements multiple AI tools with different FDA classifications. How should liability allocation differ across these tools?

    1. The hospital should assume all liability for higher-classification tools
    2. It should not differ—FDA classification has no bearing on liability
    3. Liability allocation should remain constant regardless of classification
    4. Higher-risk FDA classifications impose greater liability on the physician

11. During malpractice litigation over an AI-assisted treatment decision, what will opposing counsel most vigorously challenge regarding documentation?

    1. Whether the vendor provided training on the tool
    2. Whether the AI tool was FDA approved
    3. Whether the physician exercised independent judgment beyond AI recommendations
    4. Whether the patient signed a consent form

12. Which statement accurately reflects what AI cannot do regarding medical liability?

    1. AI cannot be used in informed consent processes
    2. AI cannot assist in documentation requirements
    3. AI cannot be held legally responsible for decisions
    4. AI cannot substitute for physician accountability

13. A physician wants to use an AI tool for a clinical purpose beyond its FDA-cleared indication. What is the appropriate approach to manage liability?

    1. Avoid using AI tools for any non-cleared indications regardless of benefit
    2. Rely on the vendor's liability clause to cover any issues
    3. Document clear clinical justification for off-label use and heightened oversight
    4. Use the tool without documentation since it's for a good clinical purpose

14. When negotiating vendor agreements for medical AI tools, what liability allocation approach appropriately protects the practice?

    1. Accepting standard vendor terms without modification
    2. Requiring the vendor to assume all liability through contract language
    3. Securing indemnification for known AI limitations while maintaining professional accountability
    4. Rejecting any vendor that does not accept full liability

15. What practical risk allocation pattern has emerged for AI in medical decisions?

    1. Liability has shifted entirely to AI vendors
    2. Physicians are no longer liable for AI-assisted decisions
    3. Liability remains with humans but documentation requirements have increased
    4. Hospitals have assumed primary liability for all AI use