From “assistant” to “autonomous”: legal liability and ethical traceability frameworks for generative AI in clinical misdiagnosis scenarios, with a special focus on paediatrics

Editorial reading guide · Trust

Who should be accountable when an AI-supported decision harms a child?

The author proposes scaling liability with a system’s practical autonomy and adding auditable clinical records of AI use. Pediatric dosing and children’s difficulty contesting errors make the allocation of responsibility directly relevant to family trust.

Study context

This theory article analyzes regulatory materials and a hypothetical clinical-error scenario; it recruited no patients, families, clinicians, or legal decision makers.

Age / stage: Neonates through adolescents are an explicit application group

Read with three questions

1. How does meaningful human review change the proposed allocation of responsibility?
2. What could an audit trail reveal that a fluent explanation cannot?
3. Which assumptions still need stakeholder testing before this framework could guide practice?

Keep the limits in view

The liability tiers and audit infrastructure are prospective proposals without empirical or stakeholder validation. The illustrative case is not an observed event, and the mainly US/EU legal framing does not establish duties in every jurisdiction.