{
  "id": 13137390,
  "title": "First real-world study of safety and quality of patient-facing AI in primary care",
  "url": "https://urgent.news/2026/10/09/first-real-world-study-of-safety-and-quality-of-patient-facing-ai-in",
  "topic": "ai",
  "section": "AI",
  "published": "2026-10-09T14:20:17.000Z",
  "source": {
    "name": "Medical Xpress",
    "slug": "medical-xpress",
    "url": "https://medicalxpress.com/news/2026-10-real-world-safety-quality-patient.html"
  },
  "original_language": "en",
  "account": "A recent study published in The Lancet has provided the first real-world evaluation of a patient-facing conversational AI system in primary care. Conducted by researchers at Beth Israel Deaconess Medical Center (BIDMC) in collaboration with Google, the study assessed the safety and quality of the Articulate Medical Intelligence Explorer (AMIE) AI chatbot used by patients after scheduling urgent appointments with BIDMC physicians.\n\nThe study enrolled 114 patients, with 98 completing both the AI interaction and their primary care appointment. Supervising physicians were present to intervene if patients exhibited signs of harm, emotional distress, or if safety concerns arose. Throughout the study, no safety-stop interventions were required, indicating that AMIE operated safely in a real-world setting.\n\nPatient interactions with AMIE were generally positive, with both patients and physician evaluators rating the quality of conversations favorably. Patients also showed improved attitudes toward AI after engaging with the chatbot. However, concerns remained regarding patient trust in the confidentiality of information shared with the system and the chatbot's honesty and trustworthiness.\n\nThe study suggests that AI tools like AMIE could enhance both patient and physician experiences, potentially improving efficiency in primary care workflows. Of the 98 completed encounters, physicians reviewed AI-generated summaries before seeing their patients in 44 cases, with 75% reporting that this preparation helped them during the visit and 57% indicating it may have influenced their clinical approach. In only one case did a provider report the interaction as potentially harmful due to anxiety caused by a possible lymphoma diagnosis.\n\nThis research marks an important step in understanding the capabilities and limitations of AI in real-world clinical settings, paving the way for future studies to determine the potential health outcomes of integrating such technologies into routine patient care.",
  "summary": "The aging population, workforce changes and patients' increasingly complex medical needs are all factors straining primary care systems around the world. Artificial intelligence tools have emerged as a potential means of extending the capabilities of health care workers, including in primary care settings. Large language models (LLMs), also known as chatbots, have performed extremely well in…",
  "key_points": [],
  "editors_take": null,
  "illustration": null,
  "coverage": {
    "outlets": 1,
    "also_reported_by": []
  },
  "ai_generated": true,
  "disclaimer": "Summaries, key points and the editor’s take are written by software from other outlets’ reporting and may contain errors — always check the linked original."
}