See the potential, address the concerns.

For many nurses, talk of artificial intelligence (AI) triggers a sinking feeling. Veteran nurses may turn a blind eye, hoping to outrun this scary technical advancement. Newer nurses may fear that it will replace them. The facts are these—AI is already here, it’s not replacing anyone (yet), and it holds enormous promise for building value around nursing care with opportunities we can’t even imagine today.

Stanford University defines AI as “the science and engineering of making intelligent machines.” For example, machine learning (ML), a form of AI, allows devices to become “smarter” over time as they take in more information. However, because all AI (including ML) is created by humans, it’s subject to the same inaccuracies, misconceptions, and biases that we are.

For example, AI could help nurses become more efficient, productive, and effective, but it doesn’t currently leverage nursing knowledge, data, and wisdom. Why is that? Simply because much nursing data hasn’t been collected and published. Volumes of medical (allopathic) data exist and have been incorporated into various forms of AI, making it valuable to physicians. Unfortunately, the lack of nursing data doesn’t preclude AI from generating “nursing content.”

This leads to the first critical underpinning of useful AI—the knowledge contained in the application, tool, or system must be based on truth and evidence. Nursing’s first and most im­portant task should be to verify the truth and evidence behind AI.

After verifying AI’s credibility, nurses will face moral and ethical dilemmas. Consider the use of a chatbot in academic coursework. Does using AI to generate written assignments facilitate learning? From a short-term perspective, ask yourself: “Did you sharpen your critical thinking skills, or did you merely satisfy a course requirement?” Now take a long view and consider what might happen when you’re in clinical practice, research, or business and are called on to leverage the knowledge you supposedly gleaned from that assignment.

In addition to truth and ethics, issues related to privacy, accountability, and reliability demand nurses’ attention. For example, was patient data analyzed with AI de-identified first? Who’s responsible for poor outcomes that result from AI systems tainted with algorithm drift? How effectively does an AI tool identify data outliers? Does an AI-based professional recruitment system target only one group of candidates (for example, women) because of bias?

The much-needed clarifying light at the end of the AI tunnel must come from boards of nursing, professional nursing organizations, and nursing certifying bodies. Just as they determine how and when simulation can be used to acquire clinical skills, they must set boundaries for AI use. With guidance from these organizations, nurses can use AI judiciously to improve patient care, manage costs, and ease the staffing shortage.

Roy L. Simpson is the assistant dean of technology management and a professor in the Nell Hodgson Woodruff School of Nursing at Emory University in Atlanta, Georgia.

American Nurse Journal. 2024; 19(1). Doi: 10.51256/ANJ012468

References

Human-Centered Artificial Intelligence. Artificial intelligence definitions. Stanford University. September 2020. hai.stanford.edu/sites/default/files/2020-09/AI-Definitions-HAI.pdf

Ortiz S. What is ChatGPT and why does it matter? ZDNet. September 15, 2023. zdnet.com/article/what-is-chatgpt-and-why-does-it-matter-heres-everything-you-need-to-know/

Pratt M. AI Accountability: Who’s responsible when AI goes wrong? TechTarget. August 19, 2021. techtarget.com/searchenterpriseai/feature/AI-accountability-Whos-responsible-when-AI-goes-wrong