Estimated reading time: 8 minutes
Nearly half of nurses are already using generative artificial intelligence at work, but only about one in five report that their organizations have formal policies or training in place, exposing a gap between how quickly AI is entering nursing practice and how healthcare systems are preparing nurses to use it.
The nursing findings from Wolters Kluwer Health’s Future Ready Healthcare Survey show 46% of nurses use GenAI at work, while only 22% report that their institutions have published policies governing its use. The same percentage said their organizations require formal training before deploying GenAI tools in nursing workflows.
For Bethany Robertson, DNP, CNM, FNAP, FAAN, Clinical Executive at Wolters Kluwer Health, that gap matters because nurses are already incorporating AI into their work while organizational standards are still developing.
“This gap tells me the integration of AI in nursing is happening the same way a lot of change happens in nursing. Something shows up that helps you get through a busy shift, so people start using it, even if the organization has not put clear rules around it yet,” Robertson told Nurse Approved. “The numbers back that up – about 46% of nurses say they are already using GenAI at work, despite just 22% having that formal guidance. That gap says the day-to-day reality is moving faster than the playbook.”
The findings reveal a complicated relationship between nurses and AI. Nurses see opportunities for the technology to reduce workload and support new staff. Still, they worry about overreliance and what happens when AI enters clinical practice without consistent standards.
Nurses See AI as a Potential Burnout Reliever
Forty-five percent of nurses surveyed said GenAI could help reduce nursing staff burnout by automating documentation, triaging routine patient questions and streamlining workflows.
Robertson sees the most realistic near-term opportunity not in replacing nursing work, but in reducing the administrative burden surrounding it.
“The most realistic near-term impact is reducing the ‘work about the work,’ not replacing the clinical work,” Robertson said. “Nurses in the survey point to things like automating documentation, summarizing medical data, and generating patient education materials. Those are practical areas where AI can reduce tension without changing the core of the nursing practice.”
An AI tool intended to improve efficiency can have the opposite effect if nurses have to spend additional time navigating it.
“If the AI solution adds steps, adds logins, or adds extra documentation, it is not a solution; it is another task,” Robertson said. “The standard should be that AI reduces time, reduces duplication, or reduces friction in a way nurses can actually feel on a shift.”
The goal, she said, should be for AI to take repetitive work off nurses’ plates rather than add another layer to their workload.
“We want AI to fade into the background. Its real purpose is for nurses to get a little more breathing room to focus on assessment, teaching, coordinating care, and spending time with patients instead of screens.”
Without AI Guardrails, Variation Can Become Risk
Nurses’ optimism is accompanied by concern about what could happen as AI becomes more deeply embedded in care. More than half of nurses surveyed (53%) said they worry GenAI could undermine decision-making skills or lead to overreliance on algorithmic outputs.
“We have to say the quiet part out loud. AI can support clinical work, but it cannot replace clinical judgment,” Robertson said.
Without shared standards, nurses may use the same technology in different ways.
“The biggest risk is the kind you do not notice right away. Two nurses can use the same tool two different ways, get two different answers, and both feel like they did the right thing,” Robertson said. “Then you start seeing it show up in the places where consistency really matters, like what ends up in the note, what a patient gets told at discharge, or what gets passed along in a handoff. That is where variation becomes risk.”
Unclear policies can also leave nurses unsure what information they can appropriately share with an AI tool. And because AI-generated responses can sound authoritative even when information is incomplete or incorrect, nurses need to critically evaluate what the technology produces.
Robertson said verification should be built into the workflow by checking AI outputs against the patient chart, unit policies, clinical decision support tools, and what the nurse is observing in the patient.
“Check it. Confirm it. Then use it.”
Clinical judgment remains the safeguard technology cannot replicate, she said.
“One thing nursing teaches you early is that what a patient says they need is not always what they truly need. Sometimes you have to listen deeper, assess carefully, and notice what is not being said. That kind of bedside judgment and presence is not something an algorithm can replace.”
AI Could Help New Nurses, but It Shouldn’t Replace Preceptors
Onboarding and training are another area where nurses see potential for AI.
Sixty-two percent of nurses said integrating AI into onboarding and training can help staff become productive faster by accelerating the transition for new graduates and transfers. Another 28% said GenAI-enhanced onboarding and training is already making new staff more productive and confident.
Robertson said AI could help a new nurse organize information, locate the appropriate policy more quickly, or work through a clinical scenario before encountering it with a patient. Some AI models can also provide feedback on skills practice or documentation.
But faster onboarding should not mean less human support.
“The key is making sure ‘faster’ does not mean ‘less supported,’” Robertson said. “But AI should never fully replace the preceptor relationship or shortcut clinical reasoning.”
Responsible AI-enabled onboarding should also teach new nurses how to verify AI-generated information, correct errors, and recognize when using the technology is inappropriate.
Formal AI Training Remains Uncommon
Despite nurses already using GenAI at work, only 22% of respondents said their organizations require formal training before deploying the tools in nursing workflows.
Robertson said education should address the situations nurses will actually encounter rather than amount to another one-time compliance module.
“At a minimum,” she said, nurses should understand what type of AI they are using and where it gets its information, what the tool can and cannot appropriately do, and how to interact with it effectively.
Training should also cover how to verify outputs against evidence-based sources and patient records, what information should never be entered into an AI system, how AI-generated material may become part of the legal medical record, how bias can affect outputs, and how nurses should escalate safety concerns.
“Training needs to be practical and tied to real scenarios, not just a one-time module,” Robertson said.
Bedside Nurses Need a Voice in AI Decisions
Policies and training alone will not determine whether AI works in nursing practice. Robertson said bedside nurses also need a meaningful role in deciding how the technology is introduced.
“In a lot of organizations, bedside nurses don’t get pulled in until after a tool is selected and the workflow is mostly designed. And then leadership is surprised when it doesn’t land well,” she said.
Waiting until implementation can mean overlooking the clinicians who understand where time is lost, where communication breaks down, and how workflows function across different shifts and patient populations.
“If nurses are going to use it, nurses need to help shape it,” Robertson said.
She said bedside nurses should help identify problems AI could address, select tools, establish governance, design pilots, and evaluate results.
Nurse input can also help determine whether a tool actually delivers value. Does it save time? Reduce documentation burden? Make handoffs safer and more consistent? Improve onboarding or the patient experience?
Robertson also sees a role for nurse champions who can bridge the gap between organizational decisions and day-to-day practice and identify problems that may not become apparent until a tool reaches the clinical workflow.
Nurses Have a Role in Protecting Patient Trust
The implications of AI extend beyond workflow and efficiency.
Because nurses spend so much time directly interacting with patients and families, Robertson said they are positioned to recognize when AI-generated information does not fit an individual patient’s circumstances or needs.
“Patients and families look to nurses for the straight answer. We are the ones at the bedside (24/7), and trust is related to proximity; we hear the questions, and we see what is landing and what is not. That is why nurses matter so much when it comes to AI.”
That includes explaining how AI is being used and recognizing when an AI-generated recommendation, education tool, or other output does not align with the patient in front of them.
“AI-supported education has to match the patient’s needs and values, not just sound polished,” Robertson said. “And if an output feels biased or simply does not line up with what you are seeing clinically, nurses need to speak up and have a clear way to escalate it.”
The Goal Isn’t More Technology
For healthcare organizations, the findings suggest the challenge is no longer simply whether to adopt AI. It is how to incorporate it into nursing practice in ways that are useful, consistent, and safe.
Robertson recommends beginning with a limited number of high-value, lower-risk applications developed with bedside nurses and nurse educators, then measuring whether they actually improve nursing work.
“Did we give time back on a shift? Did we reduce documentation burden? Did handoffs get safer and more consistent? Did onboarding improve without cutting support? If the answer is yes, scale it. If the answer is no, fix it or stop.”
As GenAI becomes part of nurses’ daily work, healthcare organizations face a clear challenge: ensuring oversight keeps pace with adoption.
Closing that gap will require more than deploying new technology. It will require practical training, clear standards, and nurses having a voice in how AI is incorporated into patient care.
“The goal is not to add technology,” Robertson said. “The goal is to free up nurses’ cognitive capacity so they can spend more time on practices that are aligned with their skill set.”
About the research: The nursing findings are based on Wolters Kluwer Health’s Future Ready Healthcare Survey, conducted by Ipsos, an independent market research firm. Respondents included physicians, nurses, pharmacists, allied health professionals, administrators, and medical librarians across the United States.
