The question everyone is asking (but few dare to say out loud)

“I’m not very comfortable with computers.” “I’m not a geek.” “I leave that to the young ones.” These phrases are heard regularly in the corridors of hospitals and medical practices when the subject of artificial intelligence comes up. And they reveal something important: for many healthcare professionals, AI comes with an implicit fear — that of not having the technical skills needed to take hold of it.

This fear is understandable. But it rests on a frequent confusion between using AI and developing AI. And this distinction changes absolutely everything.

The short answer to the question in the title? No, it is not necessary to be good at IT to use AI in healthcare. But this answer deserves to be developed, nuanced and above all explained.

The fundamental confusion: coding versus using

Let us start by clarifying an essential distinction that is not made often enough.

Developing an AI	Using an AI
Designing the algorithms	Dictating a report to Nabla and validating it
Choosing neural network architectures	Interpreting the automatic annotation of a scan (Gleamer)
Training models on data	Querying a conversational assistant to prepare a patient response
Writing code in Python or R	Reading and understanding a risk score generated by an algorithm
Manipulating libraries such as TensorFlow or PyTorch	—
Years of specialised training in mathematics and programming	—

_The difference is exactly the same as between driving a car and designing a combustion engine. Millions of people drive every day without knowing how direct injection works._

What matters for driving is understanding the rules of the road, knowing how to read a dashboard, knowing the limits of your vehicle and adapting your behaviour to conditions. Not mastering fluid mechanics. The same goes for AI in healthcare.

What you already know and which applies directly

Here is something that healthcare professionals often underestimate: their clinical training has already prepared them for several of the most important cognitive skills for using AI well.

Probabilistic reasoning

A doctor who interprets a diagnostic test knows that the positive predictive value of an examination depends on the prevalence of the disease in the tested population. They know that a very sensitive test produces false positives, and that a very specific test can miss real cases. This Bayesian reasoning is exactly the one to mobilise to interpret the outputs of an AI algorithm.

When an AI system indicates that a pulmonary nodule has an 87% probability of being malignant, a trained clinician instinctively knows that this information must be cross-referenced with the patient’s clinical context, their risk factors, the quality of the image, and the known limits of the algorithm. This is medical reasoning, not IT.

Critical thinking when faced with data

Healthcare professionals are trained to evaluate the quality of a study, to identify the biases of a publication, to distinguish correlation from causation. These same skills apply directly to the critical evaluation of an AI tool: on what data was it trained? What is its real performance in the population I treat? What are its documented limits?

Managing uncertainty

Making a clinical decision in a situation of uncertainty is a central skill of care. AI produces probabilities, scores, suggestions; never certainties. Healthcare professionals are naturally equipped to work with this type of information and to assume responsibility for the final decision.

What is useful to learn (without any coding)

If the technical skills of programming are not necessary, a certain foundation of understanding is, on the other hand, useful; even indispensable; for using AI in healthcare in an informed, safe and effective way. Here are its main components.

Understanding what an AI model is and how it learns

A healthcare professional does not need to know how to program a neural network. But they have an interest in understanding what a machine learning algorithm is in broad terms: that it learns from historical examples, that it can reproduce the biases present in this data, that it performs less well on populations different from those on which it was trained.

This high-level understanding is enough to generate the right reflexes: checking a tool’s training population before trusting it on your own patients, not extrapolating the performance advertised in a clinical study to a different context.

Knowing how to read an algorithm’s performance metrics

Sensitivity, specificity, positive predictive value, area under the ROC curve, F1 score… Some of these metrics are already familiar to you. Others less so. Knowing how to read and interpret them allows you to objectively evaluate an AI tool presented by a manufacturer or described in a publication, and to have a realistic view of it rather than blind confidence or intuitive rejection.

Understanding the challenges of confidentiality and health data

Health data is sensitive data. Using an AI tool implies knowing where your patients’ data goes, how it is protected, whether the processing complies with the GDPR and with the security framework for health information systems (PGSSI-S). This is not IT in the technical sense: it is professional responsibility.

Mastering generative AI tools for practical use

Conversational artificial intelligences; ChatGPT, Claude, Gemini, Mistral; are already used by many healthcare professionals to summarise information, prepare responses to patients, or write faster. Using them effectively requires no programming skills: it does, on the other hand, require knowing how to formulate clear instructions (what is called prompting), verifying the information produced, and knowing the cases where these tools are not reliable.

The real obstacle: it’s not technique, it’s confidence

Studies on the adoption of technologies in healthcare consistently show that the main obstacle to the use of digital tools and AI by healthcare professionals is not a lack of technical skills; it is a lack of self-confidence when faced with technology, combined with a lack of time to train.

This lack of confidence is often amplified by:

• The stereotypical representation of AI in the media, which presents it either as an ultra-complex technology reserved for Silicon Valley engineers, or as an existential danger. These two narratives contribute to distancing it from those who could benefit from it most.
• The absence of initial training on digital technologies in healthcare curricula. Many practising healthcare professionals learned their profession in an environment where advanced digital tools did not yet exist, or were not taught. This is not a lack of ability: it is a lack of exposure.
• The fear of making a mistake in a professional context where error can have consequences for patients. This caution is a precious clinical quality, but it must not prevent learning in a secure framework.

Concrete examples of healthcare professionals who use AI without IT skills

To make this tangible, here are three representative profiles of what is already happening on the ground.

Marie, GP in a rural area, 52

She has never learned to code. For six months she has been using an AI transcription assistant to generate her consultation reports. It took her two weeks to become comfortable with the tool, mainly to learn how to correct errors in local medical terminology. Today, she saves 45 minutes a day and says it has changed her professional quality of life.

Karim, coordinating nurse in a care home, 38

He uses a monitoring dashboard based on AI algorithms that aggregates his residents’ vital signs and generates alerts in the event of an anomaly. He has no idea how the algorithm works in detail. What he knows is how to interpret the alerts, when to take them seriously, when to contextualise them with his clinical observation. This is augmented nursing, not IT.

Sophie, hospital pharmacist, 44

She took a short two-day training on AI in clinical pharmacy. She now uses an AI-assisted medication reconciliation tool and takes part in the steering committee of the deployment project in her establishment. She has not coded a single line, but she has become a recognised reference on this subject in her hospital.

These three people have in common having made the choice to train; not to become engineers, but to understand enough to use AI with discernment.

What training really brings

A good AI training for healthcare professionals does not look like an IT course. It looks like what you already do well: learning in a clinical context, from concrete cases, with trainers who know your field reality.

It brings you:

• Demystification. Understanding that AI is not magic, that it has well-documented strengths and blind spots, that the results it produces are interpreted — not applied blindly. This understanding is enough to move from irrational mistrust to critical confidence.
• Mastery of everyday tools. Knowing how to effectively use a conversational AI assistant, a transcription tool, a predictive monitoring system. These are practical skills that are acquired quickly with suitable teaching.
• The ethical and regulatory framework. Knowing what an AI tool is authorised to do in your professional context, what your responsibilities are as a user, how to protect your patients. This framework is indispensable for practising with peace of mind.
• The ability to take part in your establishment’s decisions. AI projects in healthcare are poorly built when clinicians are absent from them. A trained healthcare professional can carry the voice of the field in a steering committee, identify real needs, evaluate the relevance of a tool, and avoid unsuitable deployments.

The skill that matters is clinical judgement

The question “do you need to be good at IT to use AI in healthcare?” rests on an inaccurate premise: that AI in healthcare is above all a matter of IT. This is not the case.

AI in healthcare is above all a matter of care. It produces information that healthcare professionals must interpret, contextualise and integrate into their clinical judgement. The added value does not come from the machine: it comes from the person who knows what to do with it.

What matters is not knowing how to code. It is understanding enough to trust the right tools, to remain vigilant about the limits, and never to let an algorithm decide in your place what is good for your patient. And that is precisely what a good training can bring you; in a few days, not several years.

Do you want to train in AI in healthcare without an IT background? Our programmes are designed by and for healthcare professionals: no technical prerequisites, use cases anchored in your clinical reality, and teaching adapted to your pace. Discover our training courses.

Sources and references

1. Topol, E. J. (2019). _Deep Medicine: How Artificial Intelligence Can Make Healthcare Human Again_. Basic Books.
2. He, J. et al. (2019). “The practical implementation of artificial intelligence technologies in medicine.” _Nature Medicine_, 25, 30–36.
3. Recht, M. P. et al. (2020). “Integrating artificial intelligence into the clinical practice of radiology: challenges and recommendations.” _European Radiology_, 30, 3576–3584.
4. Gong, B. et al. (2020). “Influence of artificial intelligence on Canadian medical students’ preference for radiology specialty.” _Academic Radiology_, 27(4), 566–577.
5. Obermeyer, Z. & Emanuel, E. J. (2016). “Predicting the Future — Big Data, Machine Learning, and Clinical Medicine.” _The New England Journal of Medicine_, 375(13), 1216–1219.
6. Agence du Numérique en Santé (ANS). General Security Policy for Health Information Systems (PGSSI-S). _esante.gouv.fr_.
7. Blease, C. et al. (2019). “Artificial intelligence and the future of primary care: exploratory qualitative study.” _Journal of Medical Internet Research_, 21(3), e12802.
8. Nabla — Product documentation and usage studies: _nabla.com_.
9. Gleamer — Clinical performance and CE marking: _gleamer.ai_.