Will Robots Replace Nurses? The Honest Score for 2026

The automation panic surrounding healthcare has produced a persistent misconception: that nurses are at risk. The data says otherwise, with unusual clarity.

Geppetto's Jobs Index score for Registered Nurse is 25 out of 100 — Low Risk. That places nursing among the least automatable occupations in the entire dataset, not by a narrow margin, but by a structural one. This article explains why the score is where it is, what robots actually do in hospitals today, and why the staffing crisis argument is the one no automation narrative bothers to engage with.

The Score: 25/100 — What It Means

The Geppetto composite for Registered Nurse is built from four weighted inputs:

Composite: 25/100 — Low Risk

The Oxford score of 0.9 out of 100 is not a rounding error. It is one of the lowest scores in the Oxford/Frey-Osborne dataset — lower than surgeons, lower than therapists, lower than social workers. The reason is structural: the Oxford methodology scores occupations on susceptibility to computerisation, and nursing fails almost every susceptibility criterion. Physical patient interaction, real-time clinical judgment, and adaptive human communication are not tasks that decompose into automatable sub-routines.

The IFR Deployment Reality score of 3/10 reflects the current state of commercial robotics in healthcare. Robots are deployed in hospitals at scale. None of them are replacing nurses. The Geppetto Robot Density score of 3/10 confirms this: of the 15+ medical robots tracked in the Geppetto directory, none are designed to replace bedside nursing functions. Every medical robot in the directory either assists surgeons, supports patient rehabilitation, or handles hospital logistics.

The McKinsey task automation rate for nursing is low across all core functions — patient assessment, medication administration, wound care, patient education, clinical documentation, and emotional support all resist automation for reasons that are qualitative and structural, not merely technical.

What Robots Actually Do in Hospitals Right Now

This is the most important section, because the existence of robots in hospitals is frequently cited as evidence that nursing is at risk. It is not. The robots in hospitals are doing something different.

Hospital Logistics — Aethon TUG and Moxi

The Aethon TUG is an autonomous mobile robot deployed in 200+ hospitals across the United States. It moves medications from the pharmacy to nursing units, delivers linens and supplies, and transports lab specimens. It does not assess patients, administer medications, interpret symptoms, or make any clinical decision. It is a delivery van that navigates hospital corridors.

Moxi, developed by Diligent Robotics and deployed in 20+ US hospitals, performs a similar function: room restocking, PPE delivery, lab sample transport. Moxi is notably designed as a nurse-supporting tool, not a nurse-replacing one. Diligent Robotics' explicit product thesis is that nurses spend 30% of their time on non-clinical logistics tasks, and that Moxi's purpose is to recover that time for patient care. This is the opposite of displacement.

Rehabilitation Robotics — Lokomat and ReWalk

The Hocoma Lokomat is a robotic exoskeleton used in neurological rehabilitation — stroke recovery, spinal cord injury, traumatic brain injury. It enables patients to complete high-repetition gait training that a therapist cannot physically deliver at the same volume. It is operated by physiotherapists and rehabilitation nurses, not instead of them. ReWalk Personal 6.0 is a wearable exoskeleton for spinal cord injury patients — a medical device used under clinical supervision, not an autonomous replacement for clinical staff.

For a direct comparison of the rehabilitation robotics category, see Hocoma Lokomat vs ReWalk Personal 6.0.

Therapeutic Companions — PARO

PARO Therapeutic Robot is a robotic seal deployed in 5,000+ care facilities globally, primarily in dementia and palliative care settings. Clinical evidence shows it reduces patient agitation, lowers cortisol levels, and decreases the need for sedative medication in some patient groups. It is not a nurse. It is a therapeutic tool, deployed under nursing supervision, that supplements human care in a narrow but well-evidenced way.

Surgical Assistance — Da Vinci

The da Vinci surgical system is frequently cited in automation discussions. It assists surgeons with minimally invasive procedures. It does not affect nursing staffing, nursing scope, or nursing function. Surgical robots are instruments that surgeons operate; they require scrub nurses, circulating nurses, and anaesthesia nursing throughout every procedure.

> "The question 'will robots replace nurses?' is the wrong question. The right question is: will robots allow nurses to spend less time on logistics and more time on patients? The answer to that question is already yes, in 200+ hospitals running autonomous delivery robots. That's not displacement. That's the job getting better." > — The Cricket

Why Nursing Is Structurally Hard to Automate

Three protections compound each other in nursing, and they are not incremental — they are definitional.

Physical unpredictability of patients. Patients move, resist, fall, deteriorate, and present in ways that no pre-programmed motion sequence can accommodate. A nurse repositioning a post-surgical patient with a drain, an IV line, and an oxygen saturation monitor attached is performing a task that requires real-time physical adaptation that no robot in commercial deployment handles reliably. Hospital robots that do manage physical patient interaction — transfer-assist devices in Japan — do so in tightly constrained, cooperative scenarios, not the full variability of acute care.

Real-time clinical judgment. A nurse is not executing a protocol. A nurse is continuously assessing a patient, integrating new information, and making judgment calls — whether to call the physician, whether a change in breathing pattern is significant, whether a patient's reported pain level matches their presentation. This is not a classification problem that a model resolves. It is an ongoing inferential process drawing on formal training, pattern recognition from years of clinical exposure, and contextual knowledge of this specific patient's history. Current AI systems do not replicate this in a deployable, accountable form.

Human empathy as core deliverable. Nursing is not incidentally human — it is constitutively human. Research on therapeutic alliance in healthcare consistently shows that patient outcomes are affected by the quality of human connection with care providers. Patients who feel cared for comply better with treatment, report pain more accurately, and recover differently. This is not sentiment. It is a clinical outcome driver. PARO's success in reducing agitation in dementia patients is evidence that robotic interaction can supplement human connection in a narrow therapeutic context. It is also evidence that a robotic seal requires 5,000 facilities of deployment to approximate what a nurse does in a 10-minute patient conversation.

What Will Change — The Accurate Forecast

Nursing will not be replaced. It will change. The distinction matters.

Administrative burden will reduce. Documentation, scheduling, supply chain management, and routine communication tasks that consume nursing time will increasingly be handled by software and logistics robots. This is already happening. It is good for nurses.

Robotic tools will multiply. Nurses will operate more robotic rehabilitation equipment, use autonomous monitoring systems, and work alongside logistics robots as a standard feature of hospital environments. Competency in robotic tools will become a routine nursing skill, the same way IV pump management and electronic health records did.

Specialist nursing roles in robotics-adjacent fields — robotic surgical nursing, rehabilitation robotics coordination, clinical AI oversight — will grow. These are high-skill extensions of nursing, not replacements for it.

The core job — clinical assessment, medication administration, patient education, emotional support, family communication, and care coordination — will not automate. Not in 2026. Not in 2036 under any plausible commercial robotics trajectory.

The Staffing Crisis Argument

No automation narrative about nursing engages seriously with this: the world does not have too many nurses. It has a catastrophic shortage of them.

The World Health Organisation projects a global shortfall of 10 million healthcare workers by 2030, with nurses comprising the majority of that deficit. In the United States, the Bureau of Labor Statistics projects registered nursing to add 177,000 positions annually through 2032 — making it one of the largest job growth categories in the economy. In Japan, which has adopted care robots more aggressively than any other country and has pioneered robotic patient transfer assistance, the robot investment is driven entirely by a nursing shortage so severe that care facilities cannot fill positions at any wage.

The policy and economic context is not robots competing with nurses for jobs. It is robots being developed and deployed because there are not enough nurses, and the problem is getting worse, not better.

This is the fact that makes the "will robots replace nurses?" question structurally incoherent. The market answer is already clear: robots are filling gaps humans are not available to fill, in a sector that will spend the next decade trying to recruit more humans into the profession.

Frequently Asked Questions

What is nursing's automation risk score on the Geppetto Jobs Index? Registered Nurse scores 25/100 on the Geppetto composite — Low Risk. The Oxford/Frey-Osborne automation susceptibility score for nursing is 0.9/100, one of the lowest in the entire occupational dataset. The IFR Deployment Reality score is 3/10, reflecting that no commercial robot replaces core nursing functions at scale.

Are there robots working in hospitals right now? Yes. Aethon TUG operates in 200+ hospitals handling logistics. Moxi is deployed in 20+ hospitals doing supply delivery and lab transport. PARO is used in 5,000+ care facilities as a therapeutic tool. None of these robots perform nursing functions. All of them support nursing teams by handling non-clinical tasks.

Could AI replace nursing judgment in the future? Clinical decision support AI is already used in hospitals — flagging deteriorating patients, suggesting medication adjustments, identifying sepsis risk. These tools support nursing judgment; they do not replace it. The accountability, physical assessment, and patient relationship components of nursing are not functions that AI can perform or that regulatory frameworks permit AI to perform autonomously in clinical settings.

Is Japan replacing nurses with robots? No. Japan has the world's most aggressive care robot adoption programme, driven by one of the world's most severe nursing shortages. Japanese care robots handle patient transfers and mobility assistance — physically demanding tasks that contribute to nursing burnout and injury — in cooperative, supervised settings. Japan is not reducing nursing headcount. It is trying to increase it while using robots to make the job physically sustainable.

What nursing tasks are most likely to be automated? Administrative and logistics tasks: documentation, supply ordering, routine communication, scheduling, medication transport. These are already partially automated. The clinical core of nursing — assessment, judgment, patient relationship, and hands-on care — is not on a credible automation trajectory under current or near-term robotics capability.

Should nurses be concerned about robots? The evidence-based answer is no. The global nursing shortage means demand is growing faster than supply in most healthcare systems. Robots entering hospitals are doing so to fill gaps humans cannot fill, not to compete with nurses for positions. Nurses who develop competency in robotic tools will be more competitive, not less.