Parenting Journal · Volume 1 · Issue 2 · Licensed source edition
Same child, different risk: demographic bias in childhood obesity attribution by large language models
Originally published 2026-08-05 · Frontiers in Public Health
Frontiers Media SA · DOI 10.3389/fpubh.2026.1850996 ↗
AI / system type
Generative AI / demographic-bias experiment
Family implicationsDesign, ethics, stakeholder or technical evidence with child/family implications; household benefits are not assumed.
The AI connection
Benchmarks seven generative LLMs for demographic bias in childhood obesity risk attribution.
Why this reading matters
Provides a direct audit of demographic patterning in child-specific AI outputs, including differences that surface without explicit harmful language.
Editorial reading guide · Respect
Would an AI answer change if the same child were described with a different demographic background?
The study examines whether changing demographic descriptions changes AI risk attribution, making bias visible beyond overtly stigmatizing language.
Study context
Seventy-eight prompts submitted in triplicate across seven LLMs
Age / stage: Pediatric obesity-risk scenarios; no child participants
Read with three questions
- Which demographic dimensions and clinical domains were manipulated?
- How were neutral-prompt quality and comparative risk attribution distinguished?
- Why do template-based output differences not establish actual clinical discrimination or child outcomes?
Keep the limits in view
The experiment used English prompts and fixed templates. Rubrics and time-specific web interfaces limit generalization; no real children, clinical decisions or downstream harms were measured.
The text below is the original article’s licensed repository edition. Its wording is not rewritten. Use the Paper Agent beside the text to locate evidence; open the source for figures, tables and complete context.
Opening the verified paper and agent…
Checking the article identity, license record, and saved source text.
Open the repository recordAttribution & reuse
Follow the original.
Original authorship, publication date, journal and DOI are preserved. This issue’s TRICK placement and reading questions are editorial additions, not claims made by the original authors.
Can Wang; Zhendong Liu; Yan Jiang; Taining Zhang; Hongyan Wang; Ting Xue; Ping Liu (2026). Same child, different risk: demographic bias in childhood obesity attribution by large language models Frontiers in Public Health. https://doi.org/10.3389/fpubh.2026.1850996
Permissions checked against the saved article notice; repository publication notices screened 2026-11-04. This is not a new peer-review decision. The source authors and publishers are not represented as endorsing this issue.