Designing Compassionate Systems: Why Nursing Must Lead in the Age of AI

by Gabriel Chow, RN, MN

Clinical Applications Columnist
Patient Care Manager,
Mount Sinai Hospital, Toronto, Ontario

Citation: Chow, G. (2026). Designing compassionate systems: Why nursing must lead in the age of AI. Guest Column. Canadian Journal of Nursing Informatics, 21(2). https://cjni.net/journal/?p=16754

Gabriel Chow, RN, MN, is a Patient Care Manager at Mount Sinai Hospital whose work explores the intersection of nursing leadership, digital health, and compassionate care. A nurse leader with a passion for systems thinking and emerging technologies, he writes about how innovation can strengthen—not replace—the human relationships at the heart of healthcare.

Gabriel completed both his undergraduate and graduate nursing education at Toronto Metropolitan University and previously served as a Patient Flow Coordinator at Trillium Health Partners. His work spans healthcare operations, implementation science, artificial intelligence (AI), and digital transformation, with a focus on helping organizations adopt new technologies in ways that are ethical, practical, and person-centred. He is the Principal Investigator of the Digital Pulse study, examining the digital readiness of nurse leaders, and is completing the Artificial Intelligence in Health Care Certificate Program.

Gabriel believes the future of nursing will not be defined by the technologies we adopt, but by the leadership we demonstrate in using them wisely. Through his writing, he challenges readers to think critically about the future of nursing and healthcare while remaining grounded in the enduring values of compassion, curiosity, and human connection.

I have spent my nursing career working where the health system’s cracks are most visible: at the bedside, across multiple areas as a float nurse, coordinating patient flow during the height of COVID-19, and now at the system level as a Patient Care Manager. In each role, I have carried responsibility not only for individual patients, but for how well the system around them functions.

These experiences shape how I view artificial intelligence (AI) and machine learning (ML). I do not see them as future abstractions. I see them as forces already shaping healthcare, often without sufficient nursing leadership. The question is no longer whether AI will enter practice, but whether nursing will help define its purpose, limits, and ethical foundations.

If we do not step decisively into this space, we risk being absent from decisions that will profoundly affect patients, providers, and the sustainability of our health system.

Healthcare strain and the limits of incremental change

My experience as a float nurse made one thing clear: inefficiency is rarely an individual failure. Nurses constantly adapt to fragmented information, misaligned workflows, and competing priorities. During COVID-19, coordinating patient flow amid surges revealed how quickly these inefficiencies escalate into safety risks when capacity is stretched and decisions must be made with incomplete data.

Canada’s healthcare system is under sustained pressure, serving an aging population with increasing complexity. Workforce shortages and moral distress are no longer episodic: they are structural. Continuing to rely on manual workarounds and reactive decision-making is no longer tenable.

AI and ML are not silver bullets. But they represent one of the few scalable opportunities to reduce cognitive burden, anticipate needs, and create space for nurses to practice with greater attention and judgment. Ignoring this potential does not preserve the status quo, it accelerates its erosion.

Why nursing hesitates and what is at stake

Skepticism toward AI is understandable. Nursing has lived through poorly implemented digital systems that increased documentation burden and disrupted workflows. Concerns about job displacement and erosion of the human connection are not unfounded.

Yet disengagement carries greater risk than participation.

Many nurses have had limited opportunities to develop digital and data literacy, making AI feel inaccessible. Past technology failures have eroded trust, and a culture that frames system design as external to nursing reinforces disengagement. Compounding this is the persistent myth that AI is neutral, when it inevitably reflects the values and assumptions of its designers.

When nurses are absent from design and governance, technology is experienced downstream, at the bedside, where its unintended consequences are absorbed by nurses and patients alike.

Nurses have long understood advocacy as a core responsibility. We intervene, protect, and uphold dignity in moments of care. Yet this advocacy is often localized, occurring within individual encounters rather than within the systems that shape those encounters at scale.

The design of healthcare systems: how information flows, how decisions are supported, and how care is coordinated profoundly influences both patient and provider experience. When nurses are absent from these conversations, we are left to compensate for system gaps through individual effort, often at the cost of increased cognitive burden and moral distress.

Building truly human-centred systems requires extending nursing advocacy beyond the bedside and into the design and governance of care delivery. Compassion, in this sense, is not only relational, but also structural.

Compassion cannot be left to individuals working within poorly designed systems. Without deliberate design, systems will default to efficiency over compassion, leaving clinicians to reconcile this tension in real time.

From reactive systems to anticipatory care

My transition into patient flow coordination and leadership expanded my understanding of how deeply patient outcomes are tied to operational decisions, data quality, and system design.

I repeatedly asked: Why do we rely on manual workflows when data already exist? Why is patient flow reactive rather than predictive? Why do nurses absorb the cognitive load of system inefficiencies?

AI and ML offer tools to address these challenges, supporting earlier discharge planning, anticipating bottlenecks, and aligning staffing with demand. However, without corresponding changes to workflow and system design, these tools risk amplifying inefficiencies rather than resolving them.

One area where this tension is already visible is in clinical decision support system (CDSS) tools within electronic health records. Clinical decision support is not the solution in itself, but a visible example of how AI and machine learning are entering clinical workflows. The central issue is not the tool, but who is present to shape how these tools are designed, implemented, and evaluated.

Decision-making on inpatient units is rarely linear. It is distributed across teams, shaped by time pressures and incomplete information. Introducing CDSS without attention to workflow risks increasing cognitive burden. Alerts may be ignored, workarounds may emerge, and intended benefits may not be realized. In many cases, these tools are evaluated based on whether alerts fire as intended, rather than whether decisions improve in practice.

Technology does not determine outcomes. Design does.

This is not hypothetical: early digital implementations have already shown how misaligned design shifts burden onto clinicians rather than supporting them.

For these tools to be effective, evaluation must extend beyond technical performance to include adoption, usability, and unintended consequences. The question is not whether tools function as intended, but whether they improve decision-making in practice, and whether nurses are meaningfully involved in shaping that outcome.

Critically, this requires nursing leadership and frontline clinicians to be present in design and governance. Without this, these tools risk becoming another layer of invisible work rather than meaningful support to clinical judgment.

Preparing nurses to lead in a digital system

If nurses are to lead in this space, preparation must extend beyond traditional boundaries. This does not mean turning every nurse into a data scientist. It does mean cultivating digital fluency, systems thinking, and interdisciplinary literacy.

Learning alongside fields such as informatics, human factors, and data ethics strengthens nursing’s ability to influence how care is designed and delivered.

Efforts to prohibit large language models (LLMs) in education are unlikely to be effective. Nurses are already using these tools. The ethical response is not avoidance, but preparation. When used intentionally, LLMs can support critique and strengthen clinical reasoning. Teaching nurses to appraise and contextualize AI reinforces professional judgment rather than replacing it.

A Canadian imperative for nursing leadership

Canada has an opportunity to model a value-driven approach to AI grounded in equity and public accountability. This vision cannot be realized without nursing leadership.

Nurses must be present in governance, procurement, and policy decisions. Without deliberate capacity-building, decisions about AI will continue to be made without the profession that understands care most intimately.

A call to nurse leaders

If nursing does not claim its place in shaping how care is designed and delivered, others will define it for us. The systems that shape care will be built, with or without nursing.

We have asked nurses to adapt to systems long enough: it is time to design systems around care.

Nursing advocacy must extend beyond individual encounters and into the systems that define care delivery. The future of compassionate, high-quality care will depend not only on the technologies we adopt, but on who is present to shape them.

Now is the moment for nursing to lead.