General Internal Medicine Physicians
Scrub through 147years of this role's history, from when it first emerged, through every wave of technology that reshaped it, to the cited projections for where it's heading next.
By early 2025, ambient AI clinical documentation tools have been deployed at a majority of large US health systems. Abridge reports deployment at over 250 health system clients including UPMC, Mayo Clinic, and Johns Hopkins. Microsoft Dragon Copilot (rebranded from Nuance DAX Copilot in March 2025) is deployed at over 750 US health systems. The JAMA 2025 multi-site RCT found that ambient scribes reduced after-hours EHR documentation by 62% with no loss of note quality. For the first time since the HITECH Act, general internal medicine has a credible technology intervention that directly addresses its most acute workforce crisis. Whether this reverses the pipeline decline remains to be seen: the structural economic argument for subspecializing (substantially higher compensation for procedural subspecialties) is unchanged by AI documentation tools.
The tools that defined the work
Select an era to see how it reshaped the work.
Stethoscope + percussion (Laennec 1816; post-Civil War clinical adoption)
Before the electrocardiograph, the X-ray, or the clinical laboratory, the internist's primary diagnostic instrument was the stethoscope, which Rene Laennec had invented in Paris in 1816 by rolling paper into a tube to listen to a patient's heart without direct contact. The binaural stethoscope reached American medicine by the 1850s. By the time Johns Hopkins opened in 1889, auscultation, percussion (tapping the chest to map resonance), and careful physical inspection were the entire toolkit of the physician of internal diseases. The clinical examination was everything: what you heard, felt, and saw at the bedside was the totality of diagnostic information available. Osler's great contribution was not new technology but a new discipline of observation, teaching residents that a careful physical examination could resolve most diagnostic questions that a subsequent autopsy would confirm. This era established the "physical diagnosis" tradition that remains the foundational skill of every internist trained today.
Work toolChanging equipment X-ray + clinical laboratory (Roentgen 1895; Einthoven EKG 1895-1924)
Wilhelm Roentgen announced X-rays in November 1895, and within months hospitals were using them to visualize bones, foreign bodies, and pulmonary lesions. Willem Einthoven's string galvanometer (the first practical electrocardiograph) recorded the first clinical ECG in 1895, with his landmark 1903 paper establishing the P-QRS-T notation still in use today; Einthoven received the Nobel Prize in 1924. These technologies transformed internal medicine from a purely bedside discipline into one that integrated objective instrumental data into clinical reasoning. The clinical laboratory extended this further: blood glucose measurement, bacteriological cultures, and serological tests for syphilis all arrived in the first decade of the 20th century. For the first time, the internist's diagnosis could be anchored in a number that the patient and referring physician could see. The downside was an early form of cognitive overload: clinicians debated throughout this era whether laboratory data should guide diagnosis or whether bedside judgment should remain primary.
Work toolChanging equipment Penicillin + cortisone + the therapeutics revolution (1940s-1960s)
Penicillin entered clinical medicine in 1941, fundamentally changing the internist's relationship to infectious disease: bacterial pneumonia, endocarditis, and syphilis, which had been leading causes of death managed with supportive care and hope, became curable. Cortisone followed in 1948 (Philip Hench, Mayo Clinic). Isoniazid arrived in 1952 to treat tuberculosis. Chlorpromazine (1952) began transforming psychiatry, which had been part of general medical practice. Oral diuretics (hydrochlorothiazide, 1958) and beta-blockers (propranolol, 1964) gave internists their first effective tools for hypertension and angina. This therapeutics revolution repositioned the internist from a careful diagnostician with limited treatment options to the manager of a rapidly expanding pharmacological arsenal. It also began the subspecialization pressure: the growing complexity of cardiac disease, gastrointestinal disease, and endocrine disease was generating knowledge that was too deep for any single generalist to hold.
Paper chartClinical notes Medicare + fee-for-service expansion (1966) and the specialist income divergence
Medicare became effective on July 1, 1966, and transformed the economics of American medicine. Because Congress included a provision allowing physicians to bill their "customary, prevailing and reasonable" fees, Medicare effectively let physicians set their own reimbursement for the first time within a government program. For internists, Medicare created a new and reliable payment stream for managing the chronic diseases of the aging population (hypertension, diabetes, heart disease, COPD) that dominated Medicare utilization. The structural problem was that fee-for-service payment in the Medicare fee schedule paid cognitive work (history-taking, diagnosis, counseling) at far lower rates than procedural work (colonoscopy, cardiac catheterization, surgery). This payment asymmetry drove the subspecialty income divergence: an internist who subspecialized in gastroenterology or cardiology could earn two to three times as much performing procedures as a general internist managing the same patient's underlying disease. The income gap widened every decade from the 1970s through the 1990s, creating the persistent general internal medicine workforce shortage.
Effect on the workReal physician net income for internists declined from 1975 to 1983-84 despite nominal wage increases, while surgical subspecialties gained. By the 1990s, fewer than 20% of internal medicine residency graduates chose to practice general internal medicine rather than subspecializing.
Work toolChanging equipment Hospitalist movement (Wachter + Goldman NEJM 1996; Society of Hospital Medicine 2003)
On August 15, 1996, Robert Wachter and Lee Goldman published "The Emerging Role of Hospitalists in the American Health Care System" in the New England Journal of Medicine, naming and formalizing a practice pattern that had been developing informally. The hospitalist, an internist who specializes entirely in inpatient care, would manage admitted patients full time rather than having community-based internists follow their patients into the hospital between office appointments. The Society of Hospital Medicine was named in April 2003. By the mid-2000s there were an estimated 44,000 hospitalists in the United States, most of them trained in internal medicine residency but now practicing under a distinct identity and billing code. The hospitalist movement had profound effects on the general internist role: it created a new subspecialty that absorbed a substantial share of internal medicine trainees, further reducing the pipeline to traditional outpatient general internal medicine, while simultaneously improving the coordination of inpatient care and allowing outpatient internists to focus entirely on clinic-based chronic disease management.
Effect on the workThe hospitalist specialty grew from near-zero in 1996 to an estimated 44,000 physicians by the mid-2000s, becoming the fastest-growing medical specialty in US history. Most hospitalists hold internal medicine board certification; their reclassification into hospital medicine reduced the apparent headcount of "general internists" even as the total internal-medicine-trained physician workforce grew.
Work toolChanging equipment HITECH Act + EHR mandates (2009): documentation burden and "pajama time"
The Health Information Technology for Economic and Clinical Health (HITECH) Act, signed in February 2009, appropriated $27 billion to incentivize the adoption of electronic health records and penalized non-adopters. Within a decade, EHR adoption in US hospitals rose from under 20% to over 95%. For internal medicine physicians, the EHR was simultaneously a clinical improvement (structured data, computerized prescribing, population health dashboards) and a catastrophic documentation burden: studies found that physicians spent two minutes at the computer for every one minute with patients, and that a typical primary care encounter generated 1.5 to 2 hours of EHR work per day outside of clinic hours. This "pajama time" -- documentation done at home in the evening on a laptop -- became the defining quality-of-life crisis of a generation of internists, driving burnout rates above 50% by the late 2010s and accelerating the workforce exodus from general internal medicine to better-compensated or less administratively intensive positions. The JAMA Internal Medicine 2018 paper "Physician Burnout in the Electronic Health Record Era" (Shanafelt, Dyrbye, et al.) identified EHR use as the strongest independent predictor of burnout among all physician specialties.
Effect on the workEHR-driven documentation burden is the single most commonly cited factor in the decline of internal medicine residency graduates choosing to practice general internal medicine. Studies in the 2010s found that approximately 80% of US internist time was spent on tasks other than direct patient care, with documentation consuming the largest non-clinical share.
Electronic recordDigital charting Ambient AI clinical documentation (Abridge, Dragon Copilot; JAMA 2025 RCT)
Ambient AI documentation tools, led by Abridge and Microsoft Dragon Copilot (formerly Nuance DAX), began mainstream clinical deployment in 2022 to 2024. These tools listen to patient-physician conversations during encounters and automatically generate SOAP notes, after-visit summaries, and specialist referral letters. The JAMA 2025 multi-site randomized controlled trial found that ambient scribes reduced after-hours EHR documentation by 62%, with physician satisfaction scores improving significantly and note quality rated equivalent or better. For general internists, this is the most consequential technological change since the EHR itself, running in the opposite direction: the same technology platform that imposed 2 to 4 hours of daily administrative burden is now being used to give that time back. Abridge received KLAS Best in KLAS for Ambient AI in both 2025 and 2026, deployed at UPMC, Mayo Clinic, and Johns Hopkins; Dragon Copilot is deployed at over 750 US health systems. The secondary AI layer -- clinical decision support tools (Glass Health, OpenEvidence) and chronic disease management agents (Hippocratic AI) -- extends the augmentation further into the diagnostic and panel management tasks that are the core of general internal medicine practice.
Effect on the workThe 62% reduction in after-hours documentation from the JAMA 2025 RCT, if sustained at scale, represents 1 to 2.5 clinical hours returned per physician per day. The Feinstein Institutes study warns of a general internal medicine physician shortage by 2032; ambient AI is being explored as a tool that could make the specialty sustainable enough to reverse the pipeline decline.
AI audit toolsPattern detection
What credible sources project
Scrub the slider past now to anchor each scenario on the scrubber. The spread is the range of futures credible sources project for this role.
What's shifting in the work right now
The historical view above shows how this role has moved. This is the present-day detail: which AI tools are picking up which tasks, where the edge still is, and the natural directions this work can grow.
What's changing in your day
Three parts of your work where AI is already doing real lifting, and what stays yours.
AI is sitting alongside you hereReview and approve AI-drafted SOAP notes and After Visit Summaries (AVS) generated by ambient documentation tools (Dragon Copilot, Abridge) after patient encounters — verifying each clinical data element against the Linked Sources conversation transcript, correcting diagnostic inaccuracies or omitted findings, and signing the finalized note as the physician of record before it is filed to the EHR.
Review and approve AI-drafted SOAP notes and After Visit Summaries (AVS) generated by ambient documentation tools (Dragon Copilot, Abridge) after patient encounters — verifying each clinical data element against the Linked Sources conversation transcript, correcting diagnostic inaccuracies or omitted findings, and signing the finalized note as the physician of record before it is filed to the EHR.[5],[10],[6]
Ambient AI now handles the drafting step for clinical notes — JAMA's 2025 multi-site RCT found ambient scribes cut "pajama time" (after-hours EHR work) by 62%. Your value shifts to expert clinical review and final attestation. Develop a fast, systematic review practice: treat AI note drafts as a first-pass transcript summary and scan for the specific errors these tools make — omitted pertinent negatives, imprecise symptom timing, and incorrect medication dosing — so your review is faster and more accurate than re-dictating from scratch.
AI is sitting alongside you hereManage prior authorization submissions and clinical appeals — completing EHR-native prior authorization workflows using Surescripts Prior Authorization Automation or Latent Health AI for qualifying medications and procedures
Manage prior authorization submissions and clinical appeals — completing EHR-native prior authorization workflows using Surescripts Prior Authorization Automation or Latent Health AI for qualifying medications and procedures; authoring clinical justification narratives for non-automated PA requests; and reviewing and signing AI-drafted medical necessity letters for insurer appeals.[11],[7]
Surescripts PA automation achieved 18-second median approval times at 18% automated approval rates — a meaningful fraction of the PA burden that consumed physician and staff time is now automated. The complex, high-stakes cases (specialty drugs, off-label use, devices) still require a physician-authored clinical narrative with specific diagnostic justification. Build fluency with your health system's PA tracking tools and develop efficient templates for the recurring PA types in your specialty — the physician who can turn around a complex PA narrative in 8 minutes rather than 30 minutes captures significant time back.
AI is sitting alongside you hereConduct AI-assisted pre-visit preparation — reviewing Notable Health AI-gathered patient-reported outcomes (PROs), pre-visit questionnaire responses, and automatically-populated intake data in the EHR before the encounter
Conduct AI-assisted pre-visit preparation — reviewing Notable Health AI-gathered patient-reported outcomes (PROs), pre-visit questionnaire responses, and automatically-populated intake data in the EHR before the encounter; identifying care gaps, abnormal values, and outstanding referral results pre-loaded by the AI; and using this structured pre-charting to focus the encounter on complex clinical judgment rather than data collection.[12]
AI pre-visit tools like Notable Health are reducing the data-collection burden of outpatient encounters significantly — physicians at Providence Health report arriving at encounters with intake already in the chart, allowing more time for the diagnostic conversation and exam. Build a protocol for efficiently reviewing AI-structured pre-visit summaries and train yourself to immediately recognize which pre-populated values are clinically actionable vs. routine — this is where the time savings materialize.
Where this role is heading
Natural next steps for someone with your foundation: not exits, evolutions.
Computer and Information Systems Managers
General internists who develop depth in EHR configuration, clinical AI governance, and health informatics are positioned for Chief Medical Information Officer (CMIO), VP of Clinical Informatics, or medical director of digital health roles. As ambient AI scribes (Dragon Copilot, Abridge), clinical decision support tools (Glass Health, Epic Cosmos), and AI patient engagement platforms (Hippocratic AI) become core clinical infrastructure, health systems urgently need physician leaders who can bridge the clinical and technology domains. CMIO base salaries range from $280,000–$420,000. Internists have a natural advantage over pure IT managers because they bring lived clinical workflow knowledge that is essential for evaluating AI tool safety and efficacy. A graduate certificate in health informatics (AMIA 10×10 program) or a formal biomedical informatics degree accelerates this path.
- · Health informatics credentials: AMIA 10×10 certificate (online) or Master of Biomedical Informatics for formal entry; ABPM Clinical Informatics board certification (fellowship pathway)
- · EHR platform expertise: Epic build certification (Epic EpicCare Ambulatory or Inpatient Analyst course) to understand configuration levers and AI module architecture
- · Clinical AI evaluation: model validation methodology, clinical safety metrics (sensitivity, specificity, alert precision rates), algorithmic bias auditing
- · HL7 FHIR interoperability: understanding health data exchange standards that underpin AI tool integrations and population health pipelines
- · Healthcare cybersecurity: HHS OCR HIPAA Security Rule compliance for AI systems, risk analysis for PHI in AI workflows
See the same long-arc view for your own profession.
Browse the directory by industry, or search by title or SOC code. New roles ship every few weeks. Every profile cites every claim.
Browse all roles