Paper

{
  "raw_xml": "<entry>\n    <id>http://arxiv.org/abs/2603.11413v1</id>\n    <title>Evaluation format, not model capability, drives triage failure in the assessment of consumer health AI</title>\n    <updated>2026-03-12T00:58:22Z</updated>\n    <link href='https://arxiv.org/abs/2603.11413v1' rel='alternate' type='text/html'/>\n    <link href='https://arxiv.org/pdf/2603.11413v1' rel='related' title='pdf' type='application/pdf'/>\n    <summary>Ramaswamy et al. reported in \\textit{Nature Medicine} that ChatGPT Health under-triages 51.6\\% of emergencies, concluding that consumer-facing AI triage poses safety risks. However, their evaluation used an exam-style protocol -- forced A/B/C/D output, knowledge suppression, and suppression of clarifying questions -- that differs fundamentally from how consumers use health chatbots. We tested five frontier LLMs (GPT-5.2, Claude Sonnet 4.6, Claude Opus 4.6, Gemini 3 Flash, Gemini 3.1 Pro) on a 17-scenario partial replication bank under constrained (exam-style, 1,275 trials) and naturalistic (patient-style messages, 850 trials) conditions, with targeted ablations and prompt-faithful checks using the authors' released prompts. Naturalistic interaction improved triage accuracy by 6.4 percentage points ($p = 0.015$). Diabetic ketoacidosis was correctly triaged in 100\\% of trials across all models and conditions. Asthma triage improved from 48\\% to 80\\%. The forced A/B/C/D format was the dominant failure mechanism: three models scored 0--24\\% with forced choice but 100\\% with free text (all $p &lt; 10^{-8}$), consistently recommending emergency care in their own words while the forced-choice format registered under-triage. Prompt-faithful checks on the authors' exact released prompts confirmed the scaffold produces model-dependent, case-dependent results. The headline under-triage rate is highly contingent on evaluation format and should not be interpreted as a stable estimate of deployed triage behavior. Valid evaluation of consumer health AI requires testing under conditions that reflect actual use.</summary>\n    <category scheme='http://arxiv.org/schemas/atom' term='cs.HC'/>\n    <category scheme='http://arxiv.org/schemas/atom' term='cs.AI'/>\n    <published>2026-03-12T00:58:22Z</published>\n    <arxiv:comment>12 pages</arxiv:comment>\n    <arxiv:primary_category term='cs.HC'/>\n    <author>\n      <name>David Fraile Navarro</name>\n    </author>\n    <author>\n      <name>Farah Magrabi</name>\n    </author>\n    <author>\n      <name>Enrico Coiera</name>\n    </author>\n  </entry>"
}