Registered nurses and AI

Nursing has bounded evidence for documentation assistance and nurse–LLM reception collaboration, with important study-design and reporting limits.

What published evidence shows

Nursing combines records, observation, medication administration, care planning, equipment and patient education.

Scope: official US occupation description

This is a description of responsibilities, not measured AI capability.

A 2025 study involved 40 South Korean nurses who first documented familiar virtual-patient scenarios conventionally, then with SmartENR AI in a training ENR. Authors observed shorter documentation time, but report inconsistent means and percentage statements. This edition withholds a normalized time-saving estimate.

Scope: fixed-order virtual-patient documentation pilot

Learning/order confounding and self-selected familiar scenarios limit inference.

Several authors were employed by the system provider. Virtual-patient performance does not establish clinical safety or workload savings across a nursing shift.

A published 2024 single-center randomized-trial abstract reports higher patient satisfaction for nurse–LLM reception collaboration than nurse-only reception among 2,164 participants. This is communication evidence for that setting.

Scope: outpatient reception; author abstract

The full paper methods were not independently inspected in this run.

Patient satisfaction does not establish bedside competence, safe medication administration or fewer nursing jobs.

Task boundaries

These are reference tasks and proposed evidence links, not verified AI capability ratings.

What remains unknown

Questions for a real workflow

A credible future assessment needs an exact care setting, authorized data handling, clinically appropriate quality review and measured correction effort. Documentation speed cannot establish safe medication administration, monitoring or independent patient-care competence.