The effectiveness of artificial intelligence models in addressing the concerns of families of children with cerebral palsy: a comparative analysis of ChatGPT, Gemini, and DeepSeek

Editorial reading guide · Trust

Which questions about cerebral palsy require more than a fluent general answer?

General cerebral-palsy explanations received better ratings than surgical advice. Mean scores favored Gemini descriptively, but between-model differences were not statistically significant, preserving uncertainty about a winner.

Study context

Two experts rated responses from three LLMs to ten selected orthopedic questions drawn from medical and cerebral-palsy information sites, with a third resolving discrepancies.

Age / stage: Children with cerebral palsy; no families enrolled

Read with three questions

1. Why did surgical indications expose weaker performance than basic explanations?
2. What does the nonsignificant model comparison prevent us from concluding?
3. Which family concerns were excluded when psychosocial questions were removed?

Keep the limits in view

The ten generic questions excluded psychosocial and nonorthopedic concerns and did not include patient-specific histories. Expert scores in this small snapshot do not establish decision quality, caregiver comprehension, or treatment benefit.