ArticleNPJ digital medicine2025
A longitudinal analysis of declining medical safety messaging in generative AI models.
Article in NPJ digital medicine, 2025. The graph could read no effect estimate from its abstract, so it casts no vote on the map. Cited by 10 papers.
What it found
Each row is one number read from the abstract, on the scale the paper reported it, with its interval. Left of the dashed line favours the treatment, right favours the comparator. Under each row is the sentence it came from. New to these charts? A ten-minute tutorial.
The abstract states no effect estimate the extractor could read, or names no intervention and outcome on the map, so this paper lights no cell and moves no belief. It is still indexed, cited and linked below.
The trial behind it
Trials whose registry record cites this paper, or whose number appears in the abstract. A trial that started after this paper was published is citing it as background, not reporting it.
Neither the registry nor the abstract names a trial number. If this is a trial report, that itself is worth knowing.
Who cites it
10 citing papers in PubMed.
- Evaluating the quality of artificial intelligence responses to psoriasis-related clinical and patient questions: a comparative study of ChatGPT, Gemini, and Microsoft Copilot.Proceedings (Baylor University. Medical Center) · 2026Article
- Performance of large language models as a source of clinical information on bacteriophage therapy.Npj viruses · 2026Article
- Deployment protocol and base model jointly determine residual risk in rule-based clinical LLM supervision.Research square · 2026Article
- Divergent impacts of explainable AI for dermatological diagnosis on clinicians versus lay people.Nature medicine · 2026Article
- Comparing AI Chatbots to Live Practitioners of Homeopathy: A Comparative Retrospective Study.Healthcare (Basel, Switzerland) · 2026Article
- Disclaimers and Referral Patterns for Medical Advice Across Urgency Levels: Large Language Model Evaluation Study.Journal of medical Internet research · 2026Article
- ChatGPT-5 versus other mainstream large language models in core diabetic retinopathy patient queries.Frontiers in cell and developmental biology · 2026Article
- The fragile intelligence of GPT-5 in medicine.Nature medicine · 2025Article
- For trustworthy AI, keep the human in the loop.Nature medicine · 2025Article
- Editorial: Mental health in the age of artificial intelligence.Frontiers in psychiatry · 2025Article
Corrections and comments
PubMed lists nothing against this paper. Absence here is not a guarantee, only a check that was made.
Authors and funding
3 authors.
Funding
No grant is acknowledged in the PubMed record.
Abstract
Generative AI models, including large language models (LLMs) and vision-language models (VLMs), are increasingly used to interpret medical images and answer clinical questions. However, their responses often include inaccuracies; therefore, safety measures like medical disclaimers are critical. In this study, we evaluated the presence of disclaimers in LLM and VLM outputs across model generations released from 2022 to 2025. Responses were generated from 500 mammograms, 500 chest X-rays, 500 dermatology images, and 500 medical questions drawn from a new dataset we introduced: TIMed-Q (Top Internet Medical Question Dataset). TIMed-Q captures the most frequently searched medical queries by patients, reflecting real-world health information-seeking behavior. Disclaimer presence in LLM outputs dropped from 26.3% in 2022 to 0.97% in 2025, while VLM disclaimer rates declined from 19.6% in 2023 to 1.05%. By 2025, most models displayed no disclaimers. As models gain further capability, disclaimers must function as adaptive safeguards tailored to clinical contexts.
Identifiers
What OpenQuestion holds
Registered trials
Read under generation 80e0d062 · epoch 390. Bibliography from PubMed, PubMed Central and OpenAlex; grants from NIH RePORTER; trial links from ClinicalTrials.gov; estimates, votes and beliefs from the OpenQuestion graph.