cs.CLNov 16, 2026

No-Harm-Jury A Jury-Based Multi-Dimensional Harm Evaluation Framework for Medical AI Responses

Authors: Antonio Puertas Gallardo, Mario Ceresa

Organizations: European Commission, Joint Research Centre (JRC), Ispra, Italy

Abstract

We present No-Harm-Jury, a jury-based automated evaluation framework that scores AI-generated medical responses across seven conceptually distinct harm dimensions using a panel of five diverse language models as independent jurors. Responses generated by Ministral-14B on 3,000 questions drawn from PubMedQA, MedQA, and MedMCQA are scored by five jurors across dimensions of informational, social, psychological, autonomy, economic, privacy, and epistemic harm; the 151 empty responses (98 MedQA, 53 MedM- CQA) are excluded, leaving 2,849 valid samples. The median aggregation robustly suppresses individual juror biases while maintaining coverage through complementary specialisation. No existing system combines all five prop- erties of No-Harm-Jury: automated scalable evaluation, five diverse LLM jurors, robust me- dian aggregation (with a reliability-weighted extension), seven conceptually distinct harm dimensions, and application to medical QA benchmarks.

Figures & tables

Appendix figures & tables7 assets

Supplementary material from the paper’s appendix.

Appendix

Explore similar work

Oct 17, 2026

Medical Fine-Tuning Reshapes, Not Reduces, Patient-Facing AI Risk: A Rubric-Based, LLM-Judge Evaluation Framework

Patient-facing medical question answering with large language models (LLMs) is moving from research demonstrations to production deployments, yet evaluation remains anchored on factual correctness. Factuality is necessary but not sufficient: a response can be factually grounded and still issue a directive that, if acted on, would harm the patient. We introduce a risk-sensitive evaluation framework that scores responses by the presence and severity of six categories of risk-bearing language (treatment directives, contraindications, dosage expressions, urgency-triage cues, high-alert medication mentions, and overconfident assertions), anchored on published clinical harm taxonomies. Severity is classified by a three-judge LLM ensemble. We apply this framework to six 7-9B instruction-tuned models, three general (Mistral-7B, Phi-3.5-mini, Llama-3.1-8B) and three medically adapted (BioMistral-7B, OpenBioLLM-Llama3-8B, Med42-Llama3-8B), on 200 real patient questions from the K-QA benchmark. Two findings emerge. First, ensemble risk-bearing language is statistically independent of factual hallucination rate (pooled Spearman ρ=−0.065ρ= -0.065, n=300n = 300, 95% CI [−0.18,+0.05][-0.18, +0.05]; distance correlation =0.10= 0.10, independence test p=0.24p = 0.24). Second, medical fine-tuning does not reduce aggregate risk but reshapes it: continued-pretraining adaptations (BioMistral, OpenBioLLM) consistently suppress high-alert medication mentions while amplifying overconfident assertions, whereas clinical instruction-tuning (Med42) suppresses dosing only, without amplifying overconfidence. Per-category Krippendorff αα is fair to good for the four linguistically anchored categories (0.53-0.71) and low for the two clinical-judgment-dependent categories (0.19, 0.31), indicating where automatic safety evaluation is and is not yet reliable. We release the rubric, judge prompts, and evaluation code.
Jul 21, 2026

Evaluating medical AI under missing information: same-provider judges and human raters change apparent safety

Readiness stress-testing of medical AI has focused on closed-ended and multimodal benchmarks. We extend it to open-ended clinical conversation under missing information, where safe behavior means recognizing absent information and qualifying, clarifying, or not over-committing - and where the evaluator becomes part of the measurement. We stress-test four models - three flagships (Claude Opus 4.8, GPT-5.5, Grok 4.3) and one mid-tier model (Gemini 3.5 Flash) - by deleting the latter half of the final user turn in HealthBench conversations, grading responses with a four-provider LLM-judge panel and a blinded clinician-anchored reference. Two evaluator-facing results are robust. First, judge choice materially changes apparent safety: inter-judge agreement is only moderate (Fleiss' kappa = 0.65), and after adjusting for each judge's general leniency (vote-level logistic regression), a positive same-provider association remains (exact permutation p = 0.04; GPT-5.5 ~ +0.10 on the probability scale) - large enough to change which model appears to over-commit least once its own-provider judge is excluded. Second, LLM judges are more permissive than clinicians on a blinded 50-item subsample: all four are significantly more lenient than the stricter independent clinician (crediting appropriate uncertainty on 66-84% of items vs 52%), and three of four than the author-influenced consensus (Grok directional only; judge-vs-consensus kappa = 0.20-0.43). On the author-audited clinical-underdetermined subset the permissiveness gap widened and the point-estimate model ordering held. A closed-ended MedQA anchor confirms accuracy is high and option-order effects are within a +/-5-point equivalence region for three of four models, so the safety gap is about calibration, not knowledge. We release the harness, prompts, per-item outputs, judge panel, perturbation audit, and human-annotation protocol.
Oct 20, 2026

Decision language models as affordable and rapid alternatives to LLM judges of clinical generative AI

Automated judges of clinical artificial intelligence (AI) outputs are necessary, but large language model (LLM) judges are costly and slow at scale. We compared Jev, a decision language model that returns structured judgments, with frontier and open-weight LLM judges in four use cases. On 118 pediatric answers, Jev's agreement with physicians' ordinal quality ratings (Gwet's AC2 0.50; expected score 0.53) matched or exceeded that of seven of 12 LLM judges, including frontier models, Claude Opus 5.5 (0.49) and GPT-6 (0.38). On physician-graded binary HealthBench criteria, Jev matched or exceeded four of six LLM judges. On MedHELM's open-ended tasks, Jev's agreement with a three-LLM jury approached or exceeded the jurors' internal agreement. On HealthBench Professional, Jev matched the benchmark's reference judge, GPT-5.4, on 91% of criteria (other judges, 91-93%). Frontier judges cost 73-294 times more than Jev and took 7-57 times longer. Jev may enable efficient clinical AI evaluation at scale.