Objective:

To examine UK general practitioners’ (GPs) adoption of ambient artificial intelligence (AI) scribes and to assess user-reported error rates, workflow impact and consent practices in primary care.

Approach:

• Survey Design: A nationwide online mixed-methods survey of GPs was conducted via Doctors.net.uk, targeting verified GPs.
• Data Collection: The survey included closed-ended items on AI scribe usage, errors, workload, and consent practices, along with demographic questions.

Key Findings:

• InAugust2025,of1003respondents,14%(n=141)reportedcurrentuseofambientAIscribes,39%(n=396)intendedtoadoptthemsoon,and46%(n=466)hadnoplanstousethem.Amongusers(n=141),80%(n=112)reportedreducedtimespentondocumentation,and70%(n=99)reportedreducedcognitiveload.55%(n=78)ratedthedocumentationqualityasbetterthanstandardnotes.32%(n=45)reportederrorsoften/always,with14%(n=20)experiencingsignificant-to-criticalimplications.Errorsweremostfrequentinmultipartyconsultations(38%),complexhistories(35%),andnon-Englishencounters(31%).63%(n=89)routinelysoughtpatientconsent,with≤10%ofpatientsdeclining.

Interpretation:

Ambient AI scribes provide efficiency gains and improved documentation quality but pose risks related to accuracy and equity.

Limitations:

• The sample may over-represent digitally engaged GPs due to voluntary response.
• Findings describe adoption patterns rather than population prevalence.

Conclusion:

Proactive regulation, consistent consent practices, and independent evaluation are needed for safe and equitable implementation of ambient AI scribes.