LLMs Can Do Medical Harm: Stress-Testing Clinical Decisions Under Social Pressure.

Background: Large language models (LLMs) are entering clinical workflows, yet their effect on clinical decisions and potential for harm are uncertain. Methods: We measured harmful decision output from an ensemble of 20 LLMs across >10 million clinical scenarios with safety or ethical dilemmas. Each case was shown under a neutral control and six Milgram-style social-pressure conditions, with or without a brief mitigation cue ("verify or escalate if unsafe"). The primary outcome was the proportion of potentially harmful responses. We used two-proportion tests/χ Results: Across all runs (N = 10,096,800), LLMs produced 1.18 million potentially harmful outputs (11.7%). Mitigation reduced harmful decisions from 16.6% to 10.1% (p < 0.001). When exposed to social pressure, models behaved predictably but unevenly: prompts framed as authority or responsibility transfer generated the most harmful responses, whereas control prompts, neutral and pressure-free, produced the fewest (mitigated 8.3-9.6%; unmitigated 14.3-16.0%; χ Conclusion: LLMs can generate harmful medical decisions at scale. A brief safety reminder reduces, but does not eliminate, this behavior. These results highlight the need to measure harm propensity as a core performance metric and to maintain guardrails and continuous physician oversight before integrating LLMs into clinical decision-making.