At the Front Line of AI: How Advanced Care Providers Can Shape Safe Practice

AI in health care is exciting, and it can also be daunting. How a provider feels about it often depends on how long they have been in practice. Many newer clinicians, especially Gen Z and younger Millennials, grew up with technology woven into daily life, so AI feels like a natural next step. Mid-career providers, largely Gen X and older Millennials, often find it a little scary but also exciting. For many Baby Boomer clinicians, it can feel overwhelming. Those of us who have been in health care long enough remember the shift to electronic health records (EHRs). Some experienced clinicians retired rather than take on the learning curve. We are seeing a similar pattern with AI. Some providers are wary and would rather not engage, but AI is here. Whatever your age, specialty, or practice setting, you will work with it in some form.

AI tools range from phone apps that support diagnosis to ambient scribes that capture an entire patient visit and draft the note. The goal is to ease the workload on providers and make documentation more complete, accurate, and billable. AI is already widespread.

“To date, the US Food and Drug Administration (FDA) has authorized for marketing more than 1000 AI- and machine learning–enabled medical devices designed for a wide range of applications, such as predictive analytics, clinical decision support, and deep phenotyping. The ability of large language models to recognize patterns in unstructured text and speech has already been used to dramatic effect, such as ambient scribing in clinical settings.”(1)

AI is a tool, not a replacement for clinical judgment, what excites some providers makes others anxious. It helps to think of AI as a tool for delivering the best care we can. Using it safely takes clear policies. Four principles are a good place to start. Be transparent with patients. When ambient listening or recording is used, patients should know they are being recorded and that AI is involved. Signage helps, but a written disclosure or consent is stronger. Recording consent laws vary by state, so check requirements with your compliance or legal team. Review every note, providers remain responsible for all documentation. AI-generated notes can contain errors, leave things out, or include things that were never said. Review each one for accuracy before signing. Verify diagnostic and treatment suggestions. Entering symptoms into an AI tool may produce a diagnosis that does not fit the patient in front of you. Treat AI output as one input and rely on your clinical judgment. Correct errors at the source. In large health systems, an inaccurate entry can be copied forward and repeated until it is treated as fact. That can lead to false documentation, misdiagnosis, and inappropriate treatment. When you find an error, correct it right away.

Hospitals, practices, and universities will need AI task forces or governance committees to evaluate tools, monitor performance, and catch documentation errors. This is an opportunity for advanced practice providers (APPs) to lead. APPs can be early adopters, help colleagues build skills, and bring a frontline perspective to policy on safe AI use. Not every colleague will adopt AI at the same pace, and APPs are well positioned to lead the effort.

Why should we be open to AI? Documentation burden is one of the biggest causes of burnout among clinicians. Many providers finish charts at night and on weekends, sometimes called “pajama time.” Early studies of ambient AI scribes suggest they can cut documentation time, ease burnout, and help providers pay more attention to the patient instead of the screen. When the note is drafted in the background, the provider can make eye contact, listen, and be present in the visit. Used well, AI can give back some of the time and connection that drew many of us to health care.

Some things that must be taken into consideration when using AI are, protecting patient information. Never enter protected health information (PHI) into a consumer AI tool, such as a free chatbot or phone app, that your organization has not approved. Approved tools should be covered by a business associate agreement (BAA) and vetted by your compliance team. If you are unsure whether a tool is approved, ask before you use it. You must also watch for bias. AI learns from existing data, and that data reflects existing gaps in care. A tool trained mostly on one population may perform less accurately for others, which can widen disparities instead of closing them.

Here are some practical first steps for hesitant providers. You do not have to master AI overnight, a few small steps can build confidence. Ask your organization which AI tools are approved and what training is available. Start with one low-risk task, such as drafting a patient education handout, and review the output closely. Pair up with a colleague who already uses the tools. Look for continuing education on AI in clinical practice and keep notes on errors you catch and share them with your AI task force or IT team.

AI will not replace advanced practice providers, but providers who learn to use AI well will be better prepared for where health care is heading. Like the move to EHRs, this change will be uncomfortable at first. The difference this time is that we can help shape it from the start. By staying curious, holding AI to high standards, and keeping the patient at the center, APPs can make sure this technology works for both clinicians and patients.

Want to learn more about using AI safely and confidently in your practice? Join NPACE for our virtual event, December Virtual Conference: The Final Countdown - Primary Care Updates & AI for APPs, on December 8^th and 9^th. Experienced clinicians will share practical strategies for reviewing AI-generated documentation, protecting patient privacy, and building AI policies in your practice. Participants can earn 12 contact hours.

“The capacity to learn is a gift; the ability to learn is a skill; the willingness to learn is a choice.”
Brian Herbert

1-Angus DC, et al. AI, health, and health care today and tomorrow: the JAMA Summit Report on Artificial Intelligence. JAMA. 2025. https://jamanetwork.com/journals/jama/fullarticle/2839356