Canonical text (the exact bytes fingerprinted)

TRUVACE RECORD VERSION
record: TRV-2026-1843
version: 1
kind: certified
reason: Certified into the record
timestamp: 2026-10-18T18:13:39.657908Z
status: published
lens: p_space
sector: health
headline: Public perspectives on artificial intelligence in healthcare in the USA: a cross-sectional survey
dek: Objective To evaluate public perspectives on artificial intelligence (AI) in healthcare in the USA and assess variation in trust, comfort and perceived fairness across demographic groups. Methods and analysis This cross-sectional survey of US adults was conducted between 21 November and 4 December 2025, via YouGov, a private public opinion research firm, using a non-probability-based online panel. The sample included 500 general population respondents and an oversample of 750 adults aged 60+. Survey weights were…
gain_title: (none)
problem_title: Only about a quarter of weighted US respondents felt comfortable with their provider using AI, and most preferred a provider over a chatbot for explaining care plans, which suggests deployment is outpacing public trust.
trace_subject: (none)
gain_reading: (none)
gain_evidence: (none)
problem_reading: Only about a quarter of weighted US respondents felt comfortable with their provider using AI, and most preferred a provider over a chatbot for explaining care plans, which suggests deployment is outpacing public trust.
problem_evidence: Only 25.4% reported being comfortable or very comfortable with their provider using AI | 88.4% reported a provider would better help them understand their care plan than a chatbot | AI implementation appears to be outpacing public trust
quick_read: A YouGov online survey fielded in late November and early December 2025 asked US adults about AI in healthcare. It covered 500 general population respondents plus an oversample of 750 people aged 60 and older, with weights applied to approximate the adult population. About a third of weighted respondents had used AI for health, but fewer were comfortable with their own provider using it.

The results matter because healthcare organizations are adopting AI while public acceptance appears to lag, and acceptance differed by education. The authors say the large neutral group suggests attitudes could change, but the sample is non-probability based, so how far the results generalize remains uncertain.
limitation: The survey used a non-probability-based online panel, weighted only to approximate the US adult population, and the cross-sectional design captures attitudes at one point in time.
tag: Evidence-backed problem
key_points: 34.0% of weighted respondents reported prior healthcare AI use, more commonly among college graduates (46.5%) than those without a high school diploma (22.9%). | 43.0% were uncomfortable with their provider using AI, 31.5% were neutral and 25.4% were comfortable. | 44.5% trusted AI much less than clinicians for health information. | Prior AI users were more favourable toward laboratory/imaging interpretation and AI-assisted referrals. | Differences by political affiliation, insurance, health status and age were not significant after FDR correction.
rundown: The study was cross-sectional and ran from 21 November to 4 December 2025. Outcomes included comfort with providers using AI, views on AI-generated second opinions, prior use, relative trust in AI versus clinicians, chatbot versus provider explanations of care plans, and perceived effects on fairness.

Analyses used survey-weighted tests with false discovery rate correction for multiple comparisons. Education level was the most consistent predictor of favourable attitudes, while other demographic differences did not hold up after correction.

The authors conclude that transparent deployment and targeted trust-building strategies are needed to support equitable adoption, noting that the one-third neutral share suggests attitudes are potentially modifiable.
sources:
- peer_reviewed | BMJ Digital Health & AI | https://doi.org/10.1136/bmjdh-2026-000183 | 2026-10-05
prev: 0000000000000000000000000000000000000000000000000000000000000000

sha256

64d726ff615d0fcc1aa964f92320829145a31c0de0a0e87fd93dae9f7b8537fc

previous

0000000000000000000000000000000000000000000000000000000000000000