Overview
Nurses are cautiously optimistic about AI in healthcare, but barriers like lack of trust, training, and poor past experiences still hinder adoption. For successful integration, healthcare organizations must use human-centered design, involve nurses in AI development, and support their autonomy, confidence, and connection to patient care.
What’s the story
A recent article by McKinsey highlights a growing openness among nurses to adopt AI tools in healthcare. Yet beneath the surface, there is still considerable discomfort—particularly from those who have never knowingly used AI or have only interacted with it through frustrating, cumbersome systems like EHRs. While the majority of nurses across age groups express a desire to see more AI integrated into their workflows, their hesitations and fears must be taken seriously.
Think of it like trying to describe a car to someone who’s only ever ridden a bike. Their frame of reference is limited. Many nurses equate AI with the same burdensome systems they’ve already wrestled with—systems that promised to help but instead slowed them down. The concept of AI as a co-pilot or assistant still feels foreign, especially when there’s limited exposure, understanding, or trust in the tools. In fact, some may be using AI without even realizing it—or may only associate it with consumer tools like ChatGPT or Siri.
Key barriers identified in the McKinsey report include:
- Concerns about accuracy
- Loss of human connection in care
- Lack of knowledge or training on how to use AI tools
These concerns aren’t new. When the healthcare industry transitioned from paper to EHRs in the early 2000s, many clinicians had never used a computer before—let alone charted on one. Nurses had to learn how to type, navigate a mouse, and operate laptops before they could even engage meaningfully with digital tools. Similar challenges resurfaced during the COVID-19 pandemic, when clinicians were asked to quickly adapt to Microsoft Teams and other digital platforms for patient care and internal collaboration. Many didn’t know how to log in, use a webcam, or access shared files.
We are at a similar tipping point with AI.
Assuming nurses will be comfortable with AI simply because it’s available is a misstep. Many haven’t had adequate exposure or training. Others may have tried it once—only to experience a “hallucination” or inaccurate output that eroded their trust. Worse, without education and thoughtful design, clinicians may fall back on workarounds, just like they did with EHRs—reverting to paper notes, spreadsheets, or disconnected systems.
What’s needed isn’t just technology—it’s change management.
Healthcare organizations must take deliberate steps to:
- Involve nurses and clinicians early in AI design and implementation
- Demonstrate how AI can augment—not replace—their autonomy
- Build literacy and confidence through training and deliberate practice
- Make the benefits clear: saving time, reducing burnout, and elevating the parts of care they’re most passionate about
Human-centered design is key. So is supporting the basic psychological needs of nurses:
- Autonomy – preserving their role in decision-making alongside AI
- Belonging – ensuring AI strengthens, not replaces, their connection to care teams and patients
- Competence – giving them the tools, training, and practice to feel capable using new technologies
Ultimately, if we want nurses to champion AI, we must design it with them, not just for them. That means acknowledging their fears, investing in their skills, and making sure the technology genuinely helps them do what they came into healthcare to do: care for people.
