Participate in the development, vetting, and implementation of these technologies.

Artificial intelligence (AI) has entered just about every conversation lately—from pop culture to the future of healthcare. Within the context of patient care, robots assist with surgery, algorithms can predict certain conditions, and we’ve leveraged AI in drug development. However, along with the great promise of AI, comes many questions, such as, Will it replace clinical jobs? Is it ethical and safe? Can AI really create efficiencies for clinicians? Given the rates of burnout and increases in documentation burden, a focus on impactful AI solutions that might decrease nonvalue-added documentation-related tasks for nurses seems like an important place to start.

About documentation burden

The 2022 U.S. Surgeon General’s advisory Addressing Health Worker Burnout reported that increased workload and nurse burnout go hand in hand. Much of nursing workload involves electronic health record (EHR) interactions—generation, review, analysis, and synthesis of patient data. Documentation burden (the extra effort required to use clinical documentation) plays a critical role in healthcare given the high turnover rates among clinicians.

When considering documentation burden, a few issues come to mind. For example, documentation burden prevents nurses from performing as knowledge workers. From a clinical perspective, some patients and others may perceive that, rather than synthesizing data and critically examining what’s happening to a patient, nurses simply serve as low-skilled data entry clerks. Operationally, documentation entry takes nurses away from the critical time that should go toward patient care and communication with other clinicians. In other words, documentation burden prevents nurses from functioning as they should in their healthcare setting.

Research shows that documentation burden significantly contributes to workload levels and related clinical burnout. As reported by the Centers for Disease Control and Prevention in 2023, burnout leads to poor health outcomes, reduced patient safety, and nurse turnover.

AI explained

In general, the use of AI, in everyday life or healthcare, isn’t new, but the launch of OpenAI’s ChatGPT in November 2022 made it a popular media topic. In 1995, John McCarthy, one of the founding fathers of the field, described AI as the science of making intelligent computer programs. It serves as a system feature, a method, and a tool. The National Institutes of Health defines AI as a feature in which machines learn to perform tasks, and the National Science Foundation notes that AI enables computers and other automated systems to perform tasks that have historically required human decision-making abilities. Other interpretations consider the user’s specific goal, such as using large volumes of data to answer a question or including elements of human reasoning but not requiring accurate modeling of human processes.

Several factors have increased AI’s popularity. For example, advancements in technology have made AI more accessible and capable, allowing for its integration into a wide range of applications—from smartphones and medical devices to transportation vehicles. In addition, AI can learn from increasing amounts of data, which impacts its performance and usefulness. More data mean more information to train AI, thus creating “smarter” tools.

Businesses and organizations also recognize the potential for AI to improve efficiencies, such as manufacturing a more consistent snack food package, helping people make better decisions as seen in online dating apps, and enhancing customer experiences (think Amazon’s recommendation engine). As a result, the demand for AI technologies and solutions has grown significantly.

However, the hottest AI topic concerns the use of generative AI. Traditional AI (or weak AI) focuses on processing data to return results in the form of analyses or predictions (simple input and output processes based on rules). We’ve used these traditional algorithms for decades. For example, in healthcare, algorithms analyze medical imaging data to assist with diagnoses, and EHR vendors have incorporated predictive algorithms for conditions such as sepsis and deterioration.

Generative AI doesn’t rely on specific rules but rather uses that same data to create something new by learning patterns and structures from the data input. Examples of generative AI include Generative Pre-trained Transformer (GPT) to generate dialogue, DALL-E 2 to generate images and art, and Soundful for music creation. In healthcare, promise exists for the application of generative AI to tasks like information retrieval, clinical summation, and note generation. All of these applications have the potential to reduce documentation burden.

AI in nursing practice

AI may sound futuristic, but nursing already uses many tools that leverage AI to streamline workflows and support documentation activities. At times, AI can seem ubiquitous and commonplace. However, its use to assist nursing practice remains in its infancy and the widespread use of these technologies has been slow. Each of the following examples employ traditional AI methods.

Voice recognition

AI-supported voice recognition software can reduce a nurse’s dependency on keyboards and the need to track down an available computer, thereby reducing digital interactions. For instance, patient rooms enabled with voice-activated systems use AI in a few ways. Software captures the sounds produced, uses algorithms to interpret what was said, and can either process the request or return relevant information back to the device. For example, it might take voice input and add it to the EHR as documentation.

A specific-use case for nursing includes a summary of voice-activated notes to create an end-of-shift handoff report. This summary also could be shared with patients and families. BayCare Health System’s use of a voice assistant in patient rooms serves as another example. This implementation focuses on giving patients control of their environment—turning down the lights, closing or opening the blinds, or adjusting the TV—to avoid this subset of call light requests and allowing nurses to focus on patient care and treatment interactions instead.

Clinical decision support

Many current applications for nursing clinical decision support (CDS) are rules based, but AI also can enhance CDS to reduce documentation burden. AI-powered CDS can offer early identification of patients at risk and provide actionable options and predictions based on large volumes of existing patient data. This supports nursing workflows by eliminating patient record searches and avoiding additional documentation. The CONCERN (Communicating Narrative Concerns Entered by RNs) study serves as one example.

CONCERN, a CDS tool, uses nurses’ knowledge-driven behaviors within the EHR to support early prediction of hospitalized patient decompensation. The presence of documentation beyond minimum requirements indicates that the nurse likely perceived a clinically significant event or observation important enough to record. This novel analytical approach can shift how we understand and leverage clinical observational skills and clinician-entered data.

Using data science approaches, Rossetti and colleagues developed and validated CONCERN via extensive user-centered methods with registered nurses, nurse leaders, advanced practice providers, physicians, and other key stakeholders (information technology analysts and leadership). Current software standards for exchanging healthcare information electronically allow for its easy integration into EHR systems.

CONCERN specifically enhances nursing care by stratifying patients based on acuity and promoting—not supplanting—interprofessional communication. First, CONCERN stratifies patients based on prior nurse documentation. Next, the AI empowers nurses with an evidence-based objective measure derived from their own documentation reflective of nursing practice. In the past, nurses didn’t have an objective severity number to reference when communicating with physicians; “being worried” typically wasn’t persuasive enough. Now, CONCERN predicts decompensation 42 hours earlier (improved lead time) and with the same level of accuracy as other similar systems.

Because CONCERN uses nursing expertise present in documentation as opposed to other early warning systems, which typically rely on late indicators of decompensation (such as labs and vital signs), this advancement significantly enhances nursing practice. In addition, it doesn’t require alterations to nursing workflow to gather pertinent information, thereby reducing additional documentation burden.

Smart robotics and automation

Robots aren’t new to healthcare. For example, a robotic surgical system offers a minimally invasive approach controlled by a surgeon. However, just like self-driving cars, we can train autonomous robots to independently perform nursing tasks without human intervention. These “learner” robots may have applications for medication delivery, patient monitoring, and mobility assistance. Delivery robots can save time by eliminating the need for manual tasks so nurses can spend more time with patients. They can navigate hospital hallways to deliver linens, food, or medications.

In 2022, ChristianaCare deployed a collaborative robot (cobot) with a $1.5 million grant from the American Nurses Foundation. The cobot receives training via imitation learning to gather equipment and supplies for nurses. However, through integration with the EHR, the cobot will take its application a step further. By leveraging existing documentation along with AI, cobots will proactively identify when a particular task (such as a patient assessment or transport to a procedure) requires completion, allowing for more efficient workflows. Other potential uses for robots include larger robotic machines to move patients and smaller interactive robots to combat loneliness for isolated, aging patients.

Target areas

As AI continues to evolve, an array of possibilities can help improve patient care while also creating nursing efficiencies. This opportunity enable us to rethink current processes, especially those that include intuitively capturing and displaying information for nurses.

Several foundational needs exist related to documentation burden and the application of AI to nursing practice. AI development should prioritize administrative processes, nurse staffing, intake workflows, and patient education when possible. (See Foundational needs.)

Administrative processes

Generative AI can automate time-consuming and error-prone administrative tasks such as capturing and reconciling charges. It also can create efficiencies and increase access to care by streamlining authorization and appointment processes. With regard to regulatory activities, such as compliance requirements or quality measures, generative AI can extract required elements from the EHR and accelerate gap assessments so clinicians can focus only on what’s missing.

Generative AI also can assist with communication and training by providing intuitive chat interfaces that highlight critical sections of the chart that require completion. Some pilot programs have begun exploring the use of AI with electronic prior authorization processes.

Nurse staffing

AI has the potential to inform nurse staffing and support the documentation burden of staffing coordinators and nurse leaders. Using historical staffing data, AI can predict needs to ensure the scheduling of adequate numbers of nurses during peak times. Similarly, AI-powered algorithms can optimize nurse scheduling by analyzing patient needs and predicted workload alongside nurse experience levels.

Some staffing platforms leverage AI to match available nurses with open shifts to support adequate staffing levels. In addition, organizations and nurse leaders can use AI to perform market comparisons using patient volume and staff preferences to inform fair and equitable pay for nurses.

Automated workflows

Patient intake processes frequently prove labor-intensive and time-consuming for both patients and nurses. AI-driven solutions can help create a more streamlined workflow that prevents recapturing the same data on every visit. AI systems can automatically collect and verify patient information from historical data and highlight only the items that need further review. This application may prove especially useful when obtaining an accurate medication list without relying on patient recall or manual input.

Intelligent automation also can provide a seamless digital experience for patients who store data on mobile apps, wish to complete intake forms online in advance of an appointment, or communicate with providers asynchronously. Technologies that leverage patient-

centered AI have the potential to decrease the documentation burden for patients and nurses, ultimately leading to better care outcomes and an improved patient experience.

Patient education

AI also can support the patient experience by enhancing the learning process and providing accessible, personalized information. Using AI, nurses and other providers can tailor educational content to the individual needs and preferences of patients, ensuring they receive materials most relevant to their condition and treatment. Such a tool can help eliminate the need to search and consolidate educational materials, leaving more time for instruction and discussion.

AI tools can distill complex medical information into understandable language for patients and help them easily find answers to questions outside of clinical settings and eliminate calling an office or clinic. Some tools incorporate AI to answer portal messages from patients. This type of technology can support both nursing care coordinators and care managers.

Future considerations

As technology continues to evolve, nurses play a critical role in evaluating and ensuring AI’s potential for the benefit of patients. Soon, nurses may find AI in their workplace that doesn’t mesh with existing workflows or meet the needs of their patients. Some may find this a challenge, but instead, consider it an opportunity to create authentic and impactful AI for patients and caregivers.

Start by participating in AI projects from the start and throughout the entire project life cycle. That includes validating the AI technology before implementation.

Strive to educate yourself and others about the practical uses of AI. We have so much to learn, and it keeps evolving. Adequate AI education requires innovative learning pathways similar to those used in other industries, such as tech companies. Nursing school curricula should incorporate AI, and workplaces should provide on-the-job training for the current nursing workforce.

Nurses have an ethical obligation to ensure unbiased and appropriate AI. It must provide the right intervention to the right patient at the right time. To accomplish this, ask about the development of algorithms, how they work in certain situations, and what data were used to train the AI.

Ultimately, in the spirit of documentation burden reduction, AI solutions must not shift work to other disciplines; rather, they should reduce the burden holistically. (See Getting started.)

Augmenting the human experience

Although it may take some time to see extensive AI implementation at the bedside, examples already exist in which AI brings efficiencies to nursing practice. Using technology to tackle the documentation burden connected to administrative processes, nurse staffing, and intake workflows serves as an important starting point. New discoveries leveraging generative AI will transform healthcare in ways not yet realized. It will enhance the quality of not just nurses’ lives, but also the lives and well-being of patients.

AI tools won’t replace human intervention and oversight, however, nor will AI tools take the place of clinical care. Nurses have the skills and knowledge to participate in this process and help those who build AI tools to understand and augment the human experience.

Kenrick D. Cato is a professor at the University of Pennsylvania in Philadelphia. Victoria L. Tiase is an assistant professor at the University of Utah in Salt Lake City.

American Nurse Journal. 2025; 20(2). Doi: 10.51256/ANJ022530

References

Cato KD, McGrow K, Rossetti SC. Transforming clinical data into wisdom. Nurs Manage. 2020;51(11):24-30. doi:10.1097/01.NUMA.0000719396.83518.d6

Clancy TR. Artificial intelligence and nursing: The future is now. J Nurs Adm. 2020;50(3):125-7. doi:10.1097/NNA.0000000000000855

Gesner E, Dykes PC, Zhang L, Gazarian P. Documentation burden in nursing and its role in clinician burnout syndrome. Appl Clin Inform. 2022;13(5):983-90. doi:10.1055/s-0042-1757157

Jean JY. How AI can transform nursing practice. NurseJournal. November 19, 2023. nursejournal.org/articles/nursing-practice-ai

McCarthy J. From here to human-level AI. Artif Intell. 2007;171(18):1174-82. doi:10.1016/j.artint.2007.10.009

Moy AJ, Schwartz JM, Chen R, et al. Measurement of clinical documentation burden among physicians and nurses using electronic health records: A scoping review. J Am Med Informat Assoc. 2021;28(5):998-1008. doi:10.1093/jamia/ocaa325

National Institute of Biomedical Imaging and Bioengineering. Artificial intelligence (AI).  nibib.nih.gov/science-education/science-topics/artificial-intelligence-ai

Nigam JAS, Barker RM, Cunningham TR, Swanson NG, Chosewood LC. Vital signs: Health worker–perceived working conditions and symptoms of poor mental health—Quality of Worklife Survey, United States, 2018–2022. MMWR Morb Mortal Wkly Rep. 2023;72(44):1197-1205. doi:10.15585/mmwr.mm7244e1

Office of Science and Technology Policy. National artificial intelligence research and development strategic plan: 2023 update. whitehouse.gov/wp-content/uploads/2023/05/National-Artificial-Intelligence-Research-and-Development-Strategic-Plan-2023-Update.pdf

Office of the Surgeon General. Health worker burnout. August 2, 2024. hhs.gov/surgeongeneral/priorities/health-worker-burnout/index.html

Olsen E. How Baycare Health System uses Amazon’s Alexa in patient rooms. mobihealthnews. April 17, 2023. mobihealthnews.com/news/qa-how-baycare-health-system-uses-amazons-alexa-patient-rooms

OpenAI. Pioneering research on the path to AGI. openai.com/research

Ronquillo CE, Peltonen LM, Pruinelli L, et al. Artificial intelligence in nursing: Priorities and opportunities from an international invitational think-tank of the Nursing and Artificial Intelligence Leadership Collaborative. J Adv Nurs. 2021;77(9):3707-17. doi:10.1111/jan.14855

Rossetti SC, Dykes PC, Knaplund C, et al. Multisite pragmatic cluster-randomized controlled trial of the CONCERN early warning system. medRxiv. 2024; preprint.

Rossetti SC, Knaplund C, Albers D, et al. Healthcare process modeling to phenotype clinician behaviors for exploiting the signal gain of clinical expertise (HPM-Expert Signals): Development and evaluation of a conceptual framework. J Am Med Informat Assoc. 2021;28(6):1242-51. doi:10.1093/jamia/ocab006

Sengstack PP, Adrian B, Boyd DL, et al. The six domains of burden: A conceptual framework to address the burden of documentation in the electronic health record. American Nursing Informatics Association. June 23, 2020.

Tiase VL, Cato KD. From artificial intelligence to augmented intelligence: Practical guidance for nurses. OJIN Online J Issues Nurs. 2021;26(3). doi:10.3912/OJIN.Vol26No03Man04

Wang P. On defining artificial intelligence. J Artif Gen Intell. 2019;10(2):1-37.

Key words: artificial intelligence, AI, documentation burden, workload, informatics