Frequent use of clinical AI tools rose to 34% in Elsevier's 2026 clinician survey
Elsevier's Clinician of the Future 2026 survey found 49% of clinicians use AI at work. Purpose-built clinical AI is gaining fast. Its frequent use rose from 22% in 2025 to 34% this year, yet 41% of nurses say their views are rarely or never sufficiently considered in important organizational decisions.
This story was produced through MarketScale. See how Healthcare teams put it to work with Executive Thought Leadership.
Key takeaways
A sharper vendor question: which evidence base grounds the tool's answers, and can a clinician see that source at the moment of use? Transparency and authoritative data top clinicians' own list of what builds trust.
Free workspace
Turn your Healthcare expertise into content.
Record interviews, organize footage, and write with AI on a free trial of the MarketScale platform for qualifying companies. No demo required, no credit card.
Half of clinicians say AI is already improving patient care. Almost as many, 49%, use it in their work. The more telling number in Elsevier's Clinician of the Future 2026 survey is a smaller one: frequent use of AI tools built specifically for clinical work rose from 22% in 2025 to 34% this year.
Earlier editions of Elsevier's annual series focused on what clinicians expected from AI and what worried them about it. The 2026 report describes AI in use at scale. Clinicians are using it for information retrieval, documentation, literature review, decision support, workflow and administrative tasks.
That is why the clinical-tool figure is the one to watch. The authors read the jump as a sign that healthcare organizations are starting to move beyond experiments with consumer-grade AI toward purpose-built tools designed for clinical environments. The rest of the survey explains what clinicians need before they trust either kind.
Patient load behind the uptake
The adoption is happening on crowded floors. Clinicians in the survey report rising patient volumes, growing administrative burdens, workforce shortages and more complex clinical cases. The headline figures on workload and fatigue set the scene.
The pressure clinicians report in 2026
Elsevier, Clinician of the Future 2026, via Healthcare in Europe
Those global averages hide wide regional spreads. The share of clinicians treating more patients than a year ago reaches 68% in the Middle East and Africa and 64% in Asia-Pacific, against 53% in North America. Exhaustion is uneven too: the share who say tiredness has impaired their ability to treat patients hits 58% in Brazil and 52% in the Middle East and Africa.
The results point to health systems that are still short of a full recovery. Staffing problems and disruption grew worse during the Covid-19 pandemic and in the years after it, and many systems have yet to bounce back.
The time-for-care rows complicate that picture. The Middle East and Africa and Asia-Pacific, which lead on rising volumes, also sit in the higher band of clinicians who say they still have enough time for good-quality care. The report reads this as resilience: clinicians are not simply overwhelmed victims of systemic problems, and many are actively seeking new approaches, technologies and workflows that help them deliver high-quality care.
For a chief nursing or medical information officer building a business case, the administrative burden is the place to start. Clinicians expect AI to cut the time they spend searching for information, documenting encounters and handling administrative requirements. The report describes the payoff as AI taking over repetitive, time-consuming tasks, so clinicians can spend more of the day on work that needs empathy, communication, ethical judgment and complex reasoning.
General chatbots versus clinical tools
General-purpose AI still carries most of the load. Of the clinicians who use AI, 56% say they rely on general tools frequently or consistently.
The authors flag a concern here. General systems may not be grounded in evidence-based medical information, and clinicians are not always using the most appropriate tools. Healthcare-specific solutions backed by trusted clinical content, the report argues, may offer greater reliability and safety. That view also lines up with what clinicians themselves say they need.
Respondents named the features that would raise their confidence in AI:
- Transparency about how recommendations are generated
- Strong evidence of safety and effectiveness
- Access to high-quality, authoritative data sources
- Easy integration into existing clinical workflows
- Functions that reduce workload instead of adding complexity
- Clear governance, accountability and oversight
Only one or two items on that list are about raw technical performance. The report draws the same conclusion: performance alone is not enough. Clinicians want systems they can understand, evaluate and fit into professional decision-making, and the authors argue that developers, healthcare institutions, regulators and publishers all share the job of building trustworthy AI.
Most clinicians in the survey expect AI to support decisions while judgment and responsibility stay with people.
The same conditions appear on the policy side in the United States. The Centers for Medicare & Medicaid Services' Clinical Super Intelligence Roadmap describes "earned and progressive adoption": staged deployment, transparent evidence, appropriate human oversight and continuous real-world monitoring, with tools held to a better-than-today standard. It also calls for stronger evidence systems so performance can be tracked after deployment.
The roadmap proposes starting with lower-risk uses and granting greater autonomy as the evidence supports it. It names the risks it is guarding against: mistakes, inaccurate outputs, uneven performance across populations, and new privacy and cybersecurity vulnerabilities. It also offers a pointed history lesson. In some cases, it says, the electronic health record moved work from paper to screens without redesigning the workflow, making care more complicated to deliver. The clinicians' demand for tools that reduce workload instead of adding complexity reads like a guard against repeating that.
Ask every AI vendor two things: which evidence sources ground the tool's answers, and can a clinician see that source at the moment of use?
Nurses are keen but rarely consulted
Physicians use AI more than nurses do, reflecting their greater exposure to decision support, diagnostic tools and medical information platforms. Nurses often have less direct access to these technologies.
Lower use does not mean less interest. Nurses in Elsevier's survey tend to be more optimistic about AI's future impact, and more positive about its ability to improve patient care, support professional autonomy and become trustworthy over time. The report concludes that enthusiasm may be held back less by professional role than by access, training and the way organizations implement the tools.
Wolters Kluwer's 2025 Future Ready Healthcare Survey, a separate study with its own sample, found nurses to be the most optimistic group about AI's positive effect on their work. It found that 53% of nurses worry about losing clinical decision-making skill through over-reliance on algorithms. It also found only 46% feel prepared to use generative AI, and just 22% say their organization has formal guidance for it.
Nurses on generative AI
Wolters Kluwer, 2025 Future Ready Healthcare Survey
Set the 77% who call generative AI important against the 46% who feel prepared, and Wolters Kluwer's readiness gap is plain. It fits Elsevier's reading that training and implementation, not appetite, are the binding constraints for nurses. The same Wolters Kluwer report found 62% of nurses say bringing AI into onboarding and training speeds up staff productivity and confidence, which suggests one practical place to start closing that gap.
The consultation gap is the sharper problem. In Elsevier's survey, 41% of nurses say their views are rarely or never sufficiently considered when their organizations make important decisions, and 19% of physicians agree that nurses are underrepresented.
The report calls this a potential obstacle to AI rollouts, because these technologies often change nursing workflows directly while nurses may not help select, evaluate or implement them. The result can be tools that miss the practical problems found at the bedside. Nurses are also the largest healthcare profession and the one working closest to patients, so they may judge a tool on its effect on communication, workflow and patient experience as well as efficiency. The authors' prescription is blunt: transformation cannot be driven solely by administrators, physicians or technology vendors.
Nelita Iuppa, associate chief nursing officer for informatics at Cleveland Clinic, made the governance case in a November 18 piece for the system's Consult QD. AI governance, she argued, cannot sit only with IT, data science or vendors. She lists new duties for nurse informaticists: AI evaluation, clinical validation, bias and equity review, model and output monitoring, automation oversight, and AI safety and literacy, with knock-on changes to job descriptions and org charts.
Iuppa also sketches what changes at the bedside. Nurses will increasingly interpret AI-generated insights, apply clinical judgment and recognize where the technology may be wrong. Informaticists, in her account, will help decide which parts of a nurse's work stay with the nurse, which can be augmented or safely automated, and where new forms of human oversight are needed. She argues they are suited to giving nursing that voice because they understand both clinical practice and what the technology can and cannot do. Read next to Elsevier's 41% figure, her piece is one health system's answer to the gap the survey measures.
Training with the AI switched off
Clinicians want AI that leaves judgment in human hands, and medical schools are working out how to keep that judgment strong in the first place. On November 4, HealthLeaders' Christopher Cheney reported that NYU Grossman School of Medicine is stressing reasoning and critical thinking as it plans around AI tools. The school pointed to the risk that dependence on AI can wear down critical thinking skills. One problem it is targeting is "never-skilling": new physicians who lean on AI from the start may never build foundational reasoning at all.
Jesse Burk-Rafel, assistant director and vice chair for research at NYU Grossman's Institute for Innovations in Medical Education, wrote a JAMA viewpoint on the challenges AI poses in medical education. His co-author was Robert Wachter, chair of UCSF's Department of Medicine. In it, they argued that AI systems with growing capabilities are already changing, at the bedside, a bet made a century ago about what counts in training.
Foundational knowledge still matters, and we are still teaching it. But we have moved our medical school curriculum from book learning to much more critical thinking, reasoning, and bedside skills., Jesse Burk-Rafel, NYU Grossman School of Medicine, to HealthLeaders
Burk-Rafel told the outlet that the school's AI CORE course, short for AI Clinical and Operational Readiness in Education, integrates AI into clinical skills training for students and residents. Learners practice both with and without AI, the course keeps periodic AI-free practice zones, and it holds back some tools until students master core skills. That design addresses the same fear that 53% of nurses voiced in the Wolters Kluwer survey, which suggests skill erosion is a shared worry across the professions, not a physician-only one.
The school also teaches with AI, on the view that AI as an educational partner can strengthen both human thinking and human-AI collaboration. Communication Compass coaches bedside communication by listening to ambient audio from a learner's conversation with a patient. GradeGPT reviews notes from clinical skills exams and gives feedback on them. Medical students built it because they felt they received too little feedback on their clinical note writing. Diagnosis AId uses real-time electronic health record data to comment on how a learner is reasoning. Burk-Rafel said the school also wants to train faculty in AI fundamentals, including AI literacy.
What clinicians expect by 2036
Expectations are high and close. Eighty percent of clinicians in Elsevier's survey expect AI to become a critical assistant within five to ten years. Within two to three years, 56% expect it to help improve patient outcomes and 51% expect better-quality consultations.
The job description clinicians have in mind is specific: rapid access to evidence, summaries of relevant information, pattern-spotting in large datasets and help with clinical decisions. Their conditions are just as specific. They want transparency, evidence, reliability and meaningful involvement in implementation decisions. Given the nurse consultation figures, that last condition may prove the hardest for many organizations to meet.
The report is careful on one point: workforce wellbeing, leadership, education and fair access to technology will decide as much as the software does. It adds professional development and organizational culture to that list, and it closes the argument plainly. AI may be a powerful enabler, it says, but it is not a substitute for broader healthcare reform.
Sources
- Clinician of the Future 2026: How doctors and nurses are adapting to intelligent systems ↗
- Adapting Medical Education to AI Is a Must. Here's Why. ↗ · HealthLeaders
- The Evolving Role of Nursing Informatics - Consult QD ↗
- The future of nursing with AI: Where the profession stands in a new era ↗
- A Clinical Super Intelligence Roadmap: Better Answers, Faster ... ↗
Your experts belong here
Every story in MarketScale Healthcare starts with a company putting its clinicians, service-line leaders, and field engineers on the record. Buyers are already reading this topic. The only question is whose experts they find.
Service-line buyers vet vendors quietly, and your clinicians become the proof they find while doing it.
About the author
The MarketScale Newsroom reports on the companies, technologies, and trends shaping 16 B2B industries. It turns primary sources and expert commentary into clear, useful coverage for the people doing the work.