As artificial intelligence (AI) becomes increasingly prevalent in our daily lives, our views on it are split, especially in nursing: Is AI a friend or a foe? Kenrick Cato, a nurse scientist and professor of informatics at Penn Nursing, helps debunk the concern that AI is impersonal or job-threatening. And there’s evidence to support his view.
Cato is co-author of a study that used Communicating Narrative Concerns Entered by RNs (CONCERN), an early warning system and AI-powered clinical decision support tool that reduced inpatient mortality by 35% and shortened hospital lengths of stay by half a day on average.
Cato’s path began in Guyana, where his mother, a dedicated nurse, instilled in him a deep sense of purpose. When they emigrated to the U.S., Cato embraced both his love for programming and coding and his mother’s passion for caregiving. Cato believes AI is an extension of nursing that can’t replace the critical human touch nurses bring to the bedside and that it works best when informed by nurses.
EDITOR’S NOTE: In the spirit of the topic, we used AI to help rephrase some of our questions in this Q&A.
Please introduce yourself.
My name is Kenrick Cato. I’m a professor of informatics at University of Pennsylvania School of Nursing and a scientist for pediatric data and analytics at Children’s Hospital of Philadelphia. I’ve been a nurse for 18 years.
What inspired your AI and informatics journey?
I’m an immigrant, born in Guyana. My mother’s a nurse. When I was a kid, a hospital sponsored her to work in the U.S. I was a geeky kid and started writing code at age 10. Later, I worked as a paramedic and software developer. A friend asked, “Have you ever heard of nursing informatics?” That led me to pursue a BSN, MS and PhD in nursing from Columbia University. For a time, I worked as a nurse at night and as a programmer in the IT department during the day. While programming in the electronic health record (EHR), I could show up to my clinical work and see it and use it. The clinicians I worked with provided me with feedback on my EHR programming work. You don't usually get that kind of instant feedback, so it was really nice.
You mentioned the critical question nurses often ask: Will AI take my job?
In my NTI session, I start by dispelling that myth. AI isn’t here to take nursing jobs. It’s already woven into nurses’ daily work. Understanding how AI works, and whether it's appropriate for a given patient, is essential. Nurses need to treat AI as a potential ally, not a foe.
How can informatics and AI improve patient outcomes?
From a patient perspective, AI, health analytics and informatics all help by identifying patients who might have better outcomes or benefit from specific treatments. It can also help identify patients who should not get a certain type of treatment and recognize high-risk patients in real time. We can be more precise on discharge orders and interventions. AI can even help offset health disparities by removing human bias.
For clinicians, AI can tackle documentation and scheduling, and make nurses’ jobs better and less burdensome. When clinicians are better supported, patient safety and quality of care improve. And so does clinician well-being.
There’s an irony in AI being impersonal yet enabling personalized care. Say more about that.
AI is highly computational; it’s numbers. One reason I say nurses will never be replaced is the human touch – the human ability to see things that are slightly different and react. AI may seem cold, but it can actually help us deliver less biased, more precise care. For instance, a patient whose first language isn't English might be treated differently, because a human may hesitate going into a room because they don’t know that they’ll be able to give the patient everything they're asking for. AI doesn't share those concerns or hesitations. If AI is used correctly, it can mitigate biases. It can digest countless data points to tailor care in ways humans alone can’t. That gets us to more precise care.
What AI applications are already being used in clinical settings? What applications are emerging?
That’s a great question. There’s a whole bunch of predictive tools that tell us who may respond poorly to treatments or who’s at risk. Emerging innovations include ambient sensors, audio/video systems that detect if a patient is slipping out of bed, monitor repositioning, or even assess skin integrity. In NICUs, video-based AI can detect pain in preemies or track their development.
AI is aiding documentation, optimizing back-office tasks like scheduling or supply management, and advancing patient communication via chatbots. Nearly everything nurses touch is becoming more efficient thanks to AI.
How might AI measure acuity?
Some AI tools are exploring acuity, but a better term is burden: how much care a patient needs. Two patients might be equally sick, but one might demand more nursing due to a lack of support at home or social needs. AI can help identify these differences, enabling better staffing decisions that account for both clinical needs and social factors. Assessing the burden, not just the acuity, of two similar patients is so variable. It’s going to take a lot of different perspectives to get there. But people are starting to work on that with AI.
What are you learning from AI-based clinical decision support projects such as theCONCERN clinical decision and early warning system?
I have two key insights:
Clinician involvement is vital. Projects such as CONCERN, an early warning system using nursing documentation patterns, thrive only when clinicians, especially bedside nurses, are central to design and development. That's not as easy as it sounds, but it's important to make that happen. And some forces and structures make that hard sometimes, but it is important.
Nursing intuition is a real and measurable expertise. When nurses say something feels off, that’s not magic; it’s expertise embedded in the data. AI can model that expertise, reinforcing its value rather than undermining it. AI helps quantify what nurses do. What we've learned is that nursing expertise actually contributes to what the patient's outcomes are going to be. When nurses view their expertise as intuition, I think it devalues their education, training and expertise.
Can you tell me more about how AI quantifies human insights?
AI translates nursing patterns, like note frequency and action patterns, into early signals that a patient might deteriorate. It captures what might otherwise be lost when handoffs occur or when outpatient visits are infrequent. By modeling nurse behavior, AI preserves critical insights we might otherwise lose.
There’s some worry about AI bias. How can we overcome it, especially in healthcare?
Bias isn’t inherently bad; it’s variations. But it must be acknowledged. We need to understand biases in our data, minimize them, and communicate where models perform well or poorly. For example: “This model works best with 56-year-old women; it might not be so great for 25-year-old women.” Being transparent lets users make informed choices. We also must study how bias connects to health disparities, especially in underfunded research areas.
Where is AI falling short now?
Today’s AI is the worst it will ever be. People sometimes think that we’ve unwrapped this gift that’s going to solve our problems. We’re not there yet. We’re only just starting. We must do good science: Use representative data, understand biases and align AI goals. If AI is built for efficiency rather than outcomes, clinicians will see it as a foe, not a tool that helps. We need to untangle why we’re building tools and what they’re being built for.
Looking ahead, what will AI look like in five years?
Post-COVID, there’s renewed focus on clinician well-being, equitable AI and outcomes – not just efficiency. In five years, I hope AI will better balance clinical workflow, back-office tasks, and meeting patients where they’re at. And we'll see more tools that support patients where they live, not just when they’re in the hospital. I think we’ll start to see more of those kinds of things in the next five years. And I think we’ll really see a positive impact.
It’s meaningful that you've dedicated your career to this field. What drives you?
I come from a family grounded in sacrificing for others. My mother was a nurse through the HIV [in New York] and leprosy [in Guyana] outbreaks – always showing courage. Nursing was never just a job for her. I wanted to combine that spirit of care with my love of technology. Now, I can help lighten nurses' burden and amplify their expertise. I feel blessed every day.
Finally, why should we see AI as a friend, and what’s the call to action for nurses?
We need more nurses doing AI work. If nurses leave it to other people to do it for us, we won’t get the tools we need. I’d love to see more clinical nurses in hospital governance, AI governance and data governance committees. They don't have to be AI experts, because bringing their clinical expertise to the conversation will ensure the tools we have are going to be the right tools for patients and for clinicians. AI won’t replace nurses, but AI can augment their work. It learns from your expertise, preserves your insights, and lifts you rather than beats you back. But that only happens if nurses are central to its creation, help shape AI tools and advocate for inclusive design. Nurses must be the backbone of AI in healthcare — because we know what patients truly need.
Do you or someone you know have a great nurse story to tell? We want to hear from you. Submit your story idea today.