AI in Medicine

Artificial intelligence is becoming part of modern medicine. For healthcare professionals, hearing about AI is no longer enough. It is essential to understand how it works, where it can be useful, its limitations, how to protect the patient, and how to apply it responsibly in practice.

MERIC Academy offers AI in medicine courses tailored for professionals in Romania, with a focus on responsible use, GDPR, the EU AI Act, patient safety, and real-world clinical applications.
Write your awesome label here.
Write your awesome label here.

· 12 hours of training
· 6 modules
· at own pace
· graduation certificate

About course

Artificial intelligence has already entered hospitals. X-ray systems, algorithms that predict the risk of sepsis, nurses who write medical letters — all are proposed, bought and implemented. And those who have to use them, challenge them or explain to patients are clinicians.
The problem is that available training usually goes in two wrong directions: either technical courses written for engineers, full of math and code, or enthusiastic presentations that promise revolutions without saying anything about boundaries.
This course is built differently. It gives you exactly how much you need to understand to make good decisions: how these systems work, what they really can, where they fail predictably, and what responsibility is your responsibility when you use them. No math, no code, no unnecessary jargon — but also without simplifications that would leave you unprepared in front of a well-prepared provider.

For whom this course is

• Doctors of any specialty, who want to understand what is proposed to them and critically evaluate the tools that arrive in the ward
• Nurses, working directly with monitoring, alert and documentation systems
• Medical students and residents who will practice their entire careers alongside these tools
• Healthcare management staff involved in procurement, governance or compliance decisions
You do not need any prior technical training. It is not written code, no equations. If you know how to interpret the sensitivity and specificity of a diagnostic test, you already have the necessary foundation.
This course is not for you if...
...seek a programming course, want to build artificial intelligence models, or you need technical training for a career in medical informatics. There are great courses for this — this is not one of them.

Write your awesome label here.
Write your awesome label here.

What you will be able to do after this course

• Critically evaluate any AI system proposed to you, with a structured list of questions that do not require technical knowledge

• Read the performance of a model correctly — sensitivity, predictive value, ROC curve, calibration — and understand why a system “with 95% accuracy” can be useless in your ward

• Recognize documented traps: inherited bias from data, learning shortcuts, derives over time, false alarms, hallucinations

• Responsibly use generative tools in your daily activity, knowing exactly what data can be entered and what needs to be checked

• Explain to patients and colleagues what they do — and what they don’t do — an algorithm, in an accessible language

• Understands the European and Romanian legal framework: Regulation on artificial intelligence, GDPR, medical devices, professional liability

Structure of the course

Module 1 · Artificial Intelligence Landscape — 20 min

What is and what is not artificial intelligence. Difference between AI, machine learning and deep learning. Why progress has only exploded in the last decade. The first real clinical examples and their limits.

Highly EnModule 2 · Data: Artificial Intelligence Foundation — 45 min

Why a model is always the mirror of its data. Quality of data in five questions. The standards that make the data circulate (FHIR, DICOM, ICD-10, SNOMED CT, LOINC). European Health Data Area. federated learning and synthetic data. And five documenting habits that change a lot, at no cost.

Module 3 · Classical machine learning — 45 min

You find that the risk scores you already use — CHA2DS2-VASc, Framingham, Wells — are machine learning. How to correctly read the performance of a model. Three famous failures, documented in literature: the paradox of asthma-pneumonia, the commercial model of sepsis that missed two-thirds of cases and the algorithm that predicted costs instead of the need for care.

For CoaModule 4 · Neural networks and deep learning — 45 min

How the computer learned to “see” medical images, explained without any formula. The historical and, more importantly, the first randomized artificial intelligence trial in mammographic screening on more than 105,000 women. Four specific ways of failure — including what happens when a system with more than 90% accuracy in the lab ends up in a clinic where you can't pull the curtain.

Module 5 · Generative artificial intelligence — 55 min

Language models, from mechanism to practical use. How to make a good request. What the two randomized studies show on doctors — with a result that surprises. Evidence of Environmental Documentation. And the risks of another nature: hallucination as a mechanism, the tendency of the model to give yourself justice and the mutual confirmation loop between you and the tool.

Module 6 · IA responsible in clinical practice — 30 min

The four layers of regulation that overlap. The real timetable for the application of the European Regulation on Artificial Intelligence, as amended, and the stage in Romania. Who is legally responsible for what — and what doesn’t change in your professional obligations. Ethics, reduced to four practical questions. And the checklist for when a system arrives in your precinct.

How it unfolds

4 hours narrated video presentations, divided into 6 modules
8 hours individual work: tests, deepening materials, application
1 final project on a topic from your own activity
Certified in the promotion of the final project
The course can be taken at your own pace, when the program allows, from any device. There are no fixed hours and you do not miss anything if you interrupt.

What you get in each module

• Narrrated video presentation with own graphics and clinical examples
• Self-assessment test, with detailed explanations for each response
• Additional materials: the studies cited, with full references, plus a glossary of the terms in the module
All materials remain available for further consultation.
Write your awesome label here.
Write your awesome label here.

Final project and certificate

The graduation certificate is obtained by submitting and promoting a final project, applied to your activity.

It's not an academic work. It is the exercise that turns the course into usable competence: analyze a real artificial intelligence system in your field — one already used in the institution, one that has been proposed to you or one that you have encountered in the literature – and evaluate it with the tools in the course. What does he predict?
 Who was it validated? What happens at the threshold actually used? What specific risks raise and what should be checked before use?
It is, at the same time, a document that you can use further in your institution.

Why this course is different

It's built on evidence, not enthusiasm. Each important claim is supported by studies published in reference journals, and the full references are in the additional materials, so you can check them.

It's honest about the limits. You’ll find out not only where artificial intelligence works, but also where it has failed spectacularly — and why. Failure cases are the most instructive parts of the course, and most have been discovered by clinicians who have asked simple questions.

It has a European and Romanian context. Not just the American regulation, but the European Regulation on artificial intelligence, GDPR, the European Health Data Area, the situation in Romania — including the current state of the national framework.

Leaves you with tools, not information. Four lists of questions, built mode with the mode, which you can also use in five years, when concrete examples in the course will be overcome
Write your awesome label here.

Recommended Courses:

• AI in Medicine for Clinicians

• Applied Artificial Intelligence in Medicine

• Generative AI for Physicians and Researchers

• EU AI Act for Medical Professionals

Created with
```html ```