JobShiftAtlas
Source-backed profile

Physician Assistants

Provide healthcare services typically performed by a physician, under the supervision of a physician. Conduct complete physicals, provide treatment, and counsel patients. May, in some cases, prescribe medication. Must graduate from an accredited educational program for physician assistants.

BLS SOC 29-1071 · O*NET-SOC 29-1071.00 · BLS 2025–2035 · O*NET 31.0

Median annual wage, 2025

$135,880

Projected change, 2025–2035

21.1%

2025–2035 annual average openings

11,500

AI exposure

Low

Editorial analysis

A second set of digital eyes, not a second clinician

Physician assistants can benefit from AI at nearly every information-heavy stage of care. A system can summarize years of records, organize laboratory trends, suggest questions based on a patient's history, prepare documentation, help draft instructions, and bring relevant clinical information into the workflow.

The temptation will be to trust the shortcut

Clinical AI becomes most dangerous when a plausible answer discourages further thinking. A recommendation may be based on incomplete records, unusual patient characteristics, or patterns that do not fit the person being examined. Physician assistants will need to use AI as another source of information without allowing it to narrow the clinical picture too early.

  • Documentation and chart preparation can become substantially more efficient.
  • Decision support may surface possible diagnoses, interactions, or overlooked follow-up.
  • Routine patient communications can be drafted or handled automatically.
  • Physical examination, procedures, diagnostic judgment, difficult conversations, and responsibility for treatment remain clinician functions.

AI could make physician assistants more productive, particularly in settings with heavy documentation burdens. The career advantage may belong to clinicians who can use that efficiency without surrendering curiosity, independent judgment, or attention to the patient.

Labor-market picture

Two measures, different questions

These measures offer two different ways to understand the occupation: BLS projections show how national employment may change over time, while relative AI exposure highlights where tasks may intersect with AI-related capabilities. Read them with the work profile and your local context to decide what deserves closer investigation.

Employment

Employment, 2025
168,900 jobs
Projected employment, 2035
204,600 jobs
Projected change, 20252035
35,700 jobs (21.1%)

Relative AI exposure

BLS category: Low. This is a relative task-exposure classification, not a forecast of employment change, automation, wages, or worker replacement.

Relative AI exposure · Low

Work profile

Representative tasks

These task statements come from O*NET 31.0. They are source facts, not labels assigned by the AI exposure measure.

  • Make tentative diagnoses and decisions about management and treatment of patients.
  • Interpret diagnostic test results for deviations from normal.
  • Prescribe therapy or medication with physician approval.
  • Obtain, compile, and record patient medical data, including health history, progress notes, and results of physical examination.
  • Examine patients to obtain information about their physical condition.
  • Administer or order diagnostic tests, such as x-ray, electrocardiogram, and laboratory tests.

Skills that matter

Skills in the source profile

Importance uses O*NET's 1–5 scale. Level uses its 0–7 scale. Select a skill to see its source definition and reported values.

Technology and tools

These technology records come from O*NET and show software or tools reported in the occupation profile. They describe work context, not a required checklist for every employer.

  • Medical software
  • Spreadsheet software
  • Office suite software
  • Electronic mail software
  • Word processing software
  • Storage media loading software
  • Video conferencing software
  • Internet browser software

Preparation

Typical entry education
Master's degree
Related experience
None
On-the-job training
None

Ask JobShift

Ask Dr. Morgan Reed about this evidence

Morgan is the JobShift Atlas analyst powered by AI. Answers draw from the evidence on this page and explain what it can tell you.

Source provenance

Labor-market and exposure facts: bls-ep-2025-35. Work profile: onet-31.0. Verify release details on the Sources page.