Canonical text (the exact bytes fingerprinted)

TRUVACE RECORD VERSION
record: TRV-2026-1692
version: 1
kind: certified
reason: Certified into the record
timestamp: 2026-10-17T10:42:35.385923Z
status: published
lens: g_space
sector: health
headline: Conversational diagnostic artificial intelligence in ambulatory primary care: a prospective feasibility study
dek: Background Artificial intelligence (AI)-based systems show promise for assisting primary care providers (PCPs) with patient care. We aimed to evaluate the safety and quality of clinical conversations of a patient-facing conversational AI system, which engaged in real-world urgent primary care appointments. Methods In this prospective, single-centre, single-arm feasibility study, English-speaking patients aged at least 18 years interacted with the Articulate Medical Intelligence Explorer (AMIE) up to 5 days befor…
gain_title: In supervised pre-visit chats with AMIE, no conversation needed a safety stop, and clinicians rated the conversations favourably on most criteria. Primary care physicians who read the transcripts often found them helpful for visit preparation.
problem_title: (none)
trace_subject: (none)
gain_reading: In supervised pre-visit chats with AMIE, no conversation needed a safety stop, and clinicians rated the conversations favourably on most criteria. Primary care physicians who read the transcripts often found them helpful for visit preparation.
gain_evidence: Zero conversation safety stops were required on the basis of prespecified criteria. | PCPs found AMIE helpful for visit preparation in 33 of 44 cases
problem_reading: (none)
problem_evidence: (none)
quick_read: Researchers tested a patient-facing conversational AI system called AMIE in a single-centre study. Patients spoke with it up to five days before an urgent primary care appointment for one complaint. Physician safety supervisors monitored every interaction and could intervene under predefined criteria, and transcripts and summaries went to the primary care providers before the visit.

The results suggest that this kind of AI-assisted history-taking can be feasible and acceptable in real clinical settings, at least under close physician supervision. The authors present it as an early step toward clinical translation. The study was a feasibility test rather than a comparison of patient outcomes, so benefits for diagnosis or care remain unproven.
limitation: This was a single-centre, single-arm feasibility study with physician supervision. The authors state that further research is needed, and the evidence covers safety, conversation quality and acceptance rather than clinical outcomes.
tag: Evidence-backed gain
key_points: From April to November 2025, 114 patients enrolled and 98 completed both the AMIE interaction and the primary care appointment. | Supervisors noted one hallucination and added clinical information in five interactions, though no safety stop was triggered. | Clinical evaluators rated AMIE favourably in 87-100% of cases across 17 criteria, and patients in 48-96% across 16 criteria. | Primary care providers said AMIE might have changed their behaviour in 25 of 44 cases where they reviewed the transcript.
rundown: The study was prospective, single-centre and single-arm. It included English-speaking adults aged at least 18 with a single-complaint urgent primary care appointment. Its primary outcomes were the number of supervised safety stops, conversation quality as judged by clinical evaluators, and patient and provider experiences measured by surveys.

Patient attitudes towards AI improved after the AMIE interaction and stayed elevated after the visit with their physician. Providers completed post-visit surveys in 60 of the 98 cases, and in 44 of those they had reviewed the AMIE transcript before the visit.

The authors say further research is needed. They describe the findings as initial feasibility evidence rather than proof of clinical effectiveness. The study was funded by Alphabet and registered at ClinicalTrials.gov (NCT06911398).
sources:
- peer_reviewed | The Lancet | https://doi.org/10.1016/s0140-6736(26)01535-7 | 2026-10-08
prev: 0000000000000000000000000000000000000000000000000000000000000000

sha256

77ee6bb9161aa265a1868b05947ec8aed74cc0a2e6833ef55782d7ac355b9c5f

previous

0000000000000000000000000000000000000000000000000000000000000000