Responsabilidad por negligencia cuando los médicos utilizan el cuestionario de IA

Las respuestas y explicaciones aparecen a medida que realiza el cuestionario.

1. ¿Qué conjunto enumera los cuatro elementos que normalmente requiere una demanda por negligencia en los Estados Unidos?
   • Intent, motive, opportunity, harm
   • Duty, breach of the standard of care, causation, damages
   • Consent, disclosure, billing, harm
   • License, specialty, error, apology
2. Según el análisis JAMA de 2019 realizado por Price, Gerke y Cohen, ¿qué escenario conlleva el mayor riesgo de responsabilidad para un médico?
   • Following an AI recommendation for standard care
   • Rejecting a correct but nonstandard AI suggestion
   • Following an AI recommendation that departs from standard care when the patient is then harmed
   • Not using AI in a specialty where no tools exist
3. ¿Por qué una desestimación indocumentada de una bandera de AI puede perjudicar a un médico en el tribunal?
   • EHR audit logs show the alert was displayed, which suggests the clinician was on notice
   • AI companies testify against doctors
   • Dismissing alerts is a crime
   • The flag automatically becomes the standard of care
4. ¿Por qué la autorización 510(k) de la FDA no es una protección completa para un médico que utiliza una herramienta de IA?
   • The FDA does not review any AI software
   • Clearance expires after one year
   • Clearance makes the FDA the defendant
   • 510(k) clearance generally does not block state negligence claims or define the standard of care
5. Un hospital implementa un modelo sin validación ni capacitación y un paciente resulta perjudicado. ¿Qué teoría podría utilizar un demandante contra el propio hospital?
   • Learned intermediary
   • Corporate negligence over how the tool was selected, validated and rolled out
   • Product recall
   • Informed refusal
6. Why are product liability claims against AI developers harder to win, according to the guide?
   • Developers are immune under federal law
   • Software never causes harm
   • Courts often treat software as a service, and the clinician stands between the tool and the patient as a learned intermediary
   • Patients cannot identify the developer
7. Which documentation practice does the guide recommend when AI informs a clinical decision?
   • Record the tool used, its output and why you agreed or disagreed based on your own assessment
   • Never mention AI in the chart
   • Copy the vendor's marketing claims into the note
   • Record only the final diagnosis
8. An ambient AI scribe inserts an exam finding that never happened, and the physician signs the note. Who is responsible for the note's accuracy?
   • The scribe vendor alone
   • el paciente
   • No one, because the error was automated
   • The physician who signed it