An Elsevier survey of more than 2,700 clinicians globally, reported in the KFF Health News morning briefing on May 14, 2026, found a consistent divide in how nurses and physicians experience both workload and AI adoption. The headline finding: 71% of nurses globally report having adequate time with patients, compared to 60% of physicians. Nurses are also more optimistic about AI use in their organizations than doctors.

The framing in most coverage is that this is good news for nursing. Read more carefully, and it's more complicated than that.

The 71% Number Is Not What It Looks Like

Let's start with the workload data. Seventy-one percent of nurses report feeling they have adequate patient time. That means 29% do not — roughly one in three nurses globally is reporting they cannot adequately care for their patients because of time constraints. That is not a sign of a healthy profession. It is a sign that two-thirds of nurses are just barely making it work, and a third are failing to make it work even on their own assessment.

The physician number (60%) is lower, but the comparison is not straightforward. Physicians carry a different cognitive and documentation load than nurses. Attending physicians in hospital settings are often expected to round on 15–20+ patients, manage consultant requests, review labs and imaging, complete orders, and document — often simultaneously. The 40% of physicians who say they don't have adequate patient time are likely concentrated in inpatient settings with the highest panel loads. The comparison doesn't tell us that nursing is better resourced than medicine; it tells us the professions are experiencing inadequate time in different ways.

Why Nurses Are More Accepting of AI

The AI acceptance divergence is real and worth examining directly. Nurses are more willing to use organizational AI tools than physicians. The Elsevier data aligns with other surveys showing nurse-side acceptance of ambient AI documentation tools (like Abridge, which deployed to 250+ health systems in 2026), clinical decision support alerts, and AI-generated care plans.

The interpretation that feels good: nurses are more adaptable and collaborative with technology, which bodes well for AI-driven quality improvement. The interpretation that's harder to hear: nurses accept AI because they're desperate for any tool that reduces the administrative load crushing them. When your shift involves hand-charting 47 medication administrations, completing OASIS assessments, fielding five call lights simultaneously, and documenting for an entire 12-hour shift, you will accept help from a machine without extensive philosophical debate about its limitations.

Physicians' skepticism tracks with a different problem: tools marketed to improve efficiency often add work instead of removing it. EHR implementation made physician documentation significantly longer, not shorter. Physicians have been through this before. Their skepticism about AI is informed, not ignorant — it's a learned response to being promised efficiency gains that didn't materialize.

What This Means for AI Deployment in Healthcare

The data suggests that healthcare systems implementing AI tools will encounter profession-level differences in adoption velocity. Nursing staff will likely engage with ambient documentation and decision support faster than physician staff. That pattern creates a risk: if AI is deployed to nursing workflows first (because nurses are more willing), and those tools aren't validated with the same rigor applied to physician-facing tools, you can generate patient safety exposures without the clinical pushback that would otherwise catch them.

AI in nursing has real clinical stakes. An ambient documentation tool that mischaracterizes a nurse's verbal assessment, a clinical decision support alert that fires on incorrect frequency, or an AI-generated care plan that doesn't account for patient-specific contraindications — these aren't abstract risks. Nurses interact with patients at the bedside continuously. Errors that propagate through AI-assisted documentation or care planning have more direct patient-safety consequences than errors in a physician's rounder notes that are reviewed before orders are placed.

The appropriate response to nurse-side AI acceptance is not to accelerate deployment without validation — it's to involve nurses in the validation process. Nurses know what works at the bedside. Their acceptance of AI tools is a resource, not permission to shortcut the implementation rigor those tools require.

The Bigger Picture

Both numbers — 71% and 60% — represent workforces under pressure and responding to that pressure differently. Nursing's higher AI acceptance reflects both adaptability and need. Physician skepticism reflects both informed caution and resistance to change. Healthcare organizations implementing AI in 2026 need to work with both realities, not assume that one profession's posture is the correct one.