Skip to main content
← Back to publications

Evidence Brief · Workforce Transformation

Kaiser Permanente: How Ambient AI Scribes Reduce Documentation Work While Physicians Retain Clinical Authority

How Kaiser Permanente scales ambient AI scribes to reduce physician documentation burden while preserving patient permission, clinical review and final human decision authority.

Download PDF
Cover of Kaiser Permanente: When AI Enters Clinical Work, What Must Remain Human?

Overview

Kaiser Permanente’s ambient AI deployment demonstrates how professional work can be automated selectively while clinical authority remains with physicians. This health and workforce-transition deep dive examines AI that listens to a patient-clinician conversation and produces a proposed note for physician review and editing, rather than making diagnoses, treatment recommendations or care decisions.

Between October 2023 and December 2024, 7,260 physicians used ambient AI across 2,576,627 encounters. The organizational evaluation estimated 15,791 aggregate documentation hours saved; this is not a direct measure of time returned to care, staffing reductions or improved clinical outcomes. Benefits varied with adoption intensity. A separate quality-assurance pilot identified omissions, speaker confusion and erroneous assumptions.

The case connects voluntary clinician adoption, patient permission, qualified review and correction to physician responsibility for the medical record and care decisions. Its Human Authority Stack helps leaders distinguish permission, review, verification, approval, override and accountability. It offers evidence of task compression and workflow redesign while leaving long-term staffing, retention, equity and patient-outcome effects unresolved.

Full abstract

Kaiser Permanente’s ambient AI deployment demonstrates how professional work can be automated selectively while clinical authority remains with physicians. This health and workforce-transition deep dive examines AI that listens to a patient-clinician conversation and produces a proposed note for physician review and editing, rather than making diagnoses, treatment recommendations or care decisions.

Between October 2023 and December 2024, 7,260 physicians in The Permanente Medical Group used ambient AI across 2,576,627 patient encounters. The organizational evaluation estimated 15,791 aggregate hours of documentation time saved from differences in workload measures. That estimate is not a direct measurement of time returned to patient care, an equivalent staffing reduction or evidence of improved clinical outcomes. Benefits varied with adoption intensity, and positive physician and patient surveys were not universal.

Human authority is built into the workflow through voluntary clinician adoption, patient permission before recording, review and correction of proposed notes, and physician responsibility for both the medical record and the care decision. A separate quality-assurance pilot across roughly 63,000 encounters identified omitted information, speaker confusion and erroneous assumptions, informing training, vendor feedback, deployment staging and ongoing monitoring.

The report develops a Human Authority Stack to distinguish permission, review, verification, approval, override, final authority and accountability. It treats later vendor-governance and innovation material as context rather than proof of the original deployment’s outcomes. For health-system leaders, workforce teams and AI-governance practitioners, the case offers evidence of task compression and workflow redesign while leaving physician staffing, long-term retention, equity and patient-outcome effects as questions requiring further research.

Key messages

  • Ambient AI performs first-pass documentation work; physicians remain responsible for reviewing the note and making clinical-care decisions.
  • The documented scale and estimated time savings establish workload effects rather than physician replacement or equivalent staffing reductions.
  • Patient permission and voluntary clinician adoption are distinct elements of authority that sit alongside physician review and editing.
  • Quality assurance found meaningful errors and omissions, making qualified review, feedback and ongoing monitoring essential to safe scale.
  • Adoption intensity and implementation context shape benefits; experience surveys must not be translated into clinical-outcome claims.
  • External ambient-scribe studies and later governance frameworks provide context, not independent validation of Kaiser Permanente’s original outcomes.

About this publication

Suggested citation

Global AI Governance and Workforce Transformation Policy Observatory. (2026). Kaiser Permanente: When AI Enters Clinical Work, What Must Remain Human? AIGW Health & Workforce Transition Deep Dive. https://gaeedu.org/publication/kaiser-permanente-ambient-ai-scribes-physician-authority

References

  1. Ambient Artificial Intelligence Scribes: Learnings after 1 Year and over 2.5 Million Uses

    NEJM Catalyst Innovations in Care Delivery · 2025. First-year organizational evaluation / commentary with EHR workload analysis and surveys

  2. Quality assurance informs large-scale use of ambient AI clinical documentation

    Permanente Medicine / Kaiser Permanente · 26 March 2025. First-party summary of NEJM AI quality-assurance case study

  3. How AI is giving physicians more time for what matters most

    Permanente Medicine · 2025. First-party workforce / physician-wellness interpretation

  4. Ambient Artificial Intelligence Scribes to Alleviate the Burden of Clinical Documentation

    NEJM Catalyst / The Permanente Medical Group · 2024. Early TPMG implementation evaluation / commentary

  5. AI scribe saves doctors an hour at the keyboard every day

    American Medical Association · 2024. Independent professional-association reporting based on TPMG rollout

More references (14)
  1. Quality Assurance during the Rapid Implementation of an AI-Assisted Clinical Documentation Support Tool

    NEJM AI · 2025. Quality-assurance case study

  2. AI-assisted notetaking gains steady support from Kaiser Permanente physicians

    Kaiser Permanente Division of Research · 1 April 2025. First-party research summary

  3. AI scribes save 15,000 hours —and restore the human side of medicine

    American Medical Association · 12 June 2025. Independent professional-association reporting

  4. Clinician Experiences With Ambient Scribe Technology to Assist With Documentation Burden and Efficiency

    JAMA Network Open · 2025. External prospective pre-post quality- improvement study

  5. Ambient AI Scribes in Clinical Practice: A Randomized Trial

    External randomized clinical study · 2025. External randomized evidence

  6. Medical Scribe and Ambient Artificial Intelligence Impact on Emergency Physician Documentation Burden and Clinical Productivity

    Annals of Emergency Medicine · 2026. External observational comparison

  7. Podcast: AI, innovation, and value- based care in medicine

    Permanente Medicine · 2025. First-party executive discussion

  8. An eye on innovation to enhance care delivery

    Permanente Medicine · 23 March 2026. First-party annual-report governance update

  9. Lessons learned from the Kaiser Permanente rollout of ambient AI scribes

    Permanente Medicine · 20 March 2025. First-party implementation reporting

  10. Kaiser Permanente. Artificial Intelligence Requirements for Vendors, Contractors and Suppliers. n.d.. Formal organizational vendor-governance requirement
  11. Building trust in health care through better AI governance

    Permanente Medicine · 28 July 2025. First-party executive governance discussion

  12. How ambient AI can help identify patient social needs

    Permanente Medicine · 27 August 2026. First-party follow-on innovation discussion

  13. Use of Ambient AI Scribes to Reduce Administrative Burden and Professional Burnout

    JAMA Network Open · 2025. External multi-clinician observational evidence

  14. Ambient AI Scribes and the Quintuple Aim: What Is Counted—and What Matters

    JAMA · 2026. Expert editorial from TPMG / Kaiser Permanente researchers

Flagship Report

Used in this research

Workforce Architecture for the AI Enterprise

Drawing on nine enterprise AI transitions, this flagship report examines how organizations redesign workflows, roles, capabilities, human authority and careers as AI moves from adoption into everyday operations.

This case is one of the enterprise transitions examined in this report.

12 October 2026