H1 title of the lesson

Welcome to this course. From now on, we will guide you through the basics of AI for healthcare professionals, so you can start using it confidently in your daily work.

Let's get started!

Even if you don’t have a technical background, you should understand the fundamentals of AI. Why? Because it will help you make the most of it in practice and in patient care.

👉 First point: AI is not just one thing; it is a set of technologies that work together.

Think of your “Science” classes in college: physics, biology, chemistry… and their subtopics. AI works the same way: it’s an umbrella with many areas and layers, each with its own purpose, tools, and complexity.

️‍🔥 Warm-up: Invisible AI in your everyday life

We invite you to do this quick exercise and see how many AI tools you already use in your daily life, sometimes without even realizing it!

To help you, here are some of the most common ones, organized by typical use.

1. First, voice assistants like Gemini, Alexa, or Siri. They help you search for information, set reminders, or schedule appointments—all using your voice.
2. Next are chatbots like ChatGPT, Gemini, Microsoft Copilot, and DeepSeek. They can answer questions, analyze data, summarize texts, and even create images or documents.
3. You may also be using predictive text and autocorrect when your phone suggests your next word or automatically fixes a typo. It uses AI to analyze your writing patterns and predict your intent.
4. Finally, streaming services and recommendations, like Netflix, Spotify, or YouTube. They all use AI to analyze your viewing or listening history and suggest new content you might like.

So, even if you hadn’t noticed before, AI is already quietly integrated into many aspects of daily life.

In this course, we’ll go one step further, exploring how AI can support your medical practice specifically—from documentation and management to patient communication.

The intelligence behind ChatGPT, Gemini, and other popular AI tools

The more you learn to use them, the more powerful they become. So let’s quickly review the basics of these tools: how do they learn, understand, and generate human language?

They are built on something called Large Language Models (LLMs).

But… wait a moment before that intimidating term scares you off! “Large Language Models” may sound overwhelming. We’ll show you only the essentials you need to easily understand this concept.

The robot in the library metaphor

A story to easily understand what LLMs are: Imagine a robot in a huge library. This robot has read millions of books, web pages, articles, and even conversations. The robot doesn’t think like a human but is incredibly good at remembering how we usually speak, write, and ask questions.

This is essentially what an LLM does. It’s a type of AI that has learned to “speak” and “understand” human language in a very sophisticated way because it has been trained on a massive amount of data.

🕵️ Acts like a detective: after reading so much, an LLM starts detecting patterns.

For example, if you ask “What’s the weather today?”, the expected answer would be about sun, rain, or clouds.

🔮“Predicts” the next word: but it doesn’t just finish your sentence, it can also generate full texts, answer questions, summarize articles, translate languages, and much more.

It’s like a super-intelligent machine generating coherent and relevant content based on your instructions. It’s like asking a chef to make risotto. It doesn’t invent it from scratch; it consults its best cookbook based on what it has learned.

In short, a LLM is a type of AI trained to understand and generate human language. It learns by analyzing enormous amounts of data (like books, websites, and articles) to understand how language works.

A clear example is how Google now presents search results. Before, Google was an “information directory,” showing you where to find answers. Now, with LLMs, it has become an “information compiler and summarizer,” giving you answers directly, conversationally, and with context.

Imagine you want to find the best treatment options for a newly diagnosed type 2 diabetes patient. Before, Google would give you a list of links to scientific articles or clinical guidelines. Now, with LLMs, it provides a summarized answer including first-line treatments, lifestyle recommendations, and even potential drug interactions, all in a clear, conversational format.

So, with new AI platforms appearing almost daily, it’s completely normal to feel a bit lost, especially if you’re not familiar with the technical side. Sound familiar?

Before creating a prompt… how to choose the right tool?

The first thing to keep in mind: what an AI can or cannot do depends on how it was trained, i.e., the type of model it’s based on. Here we break down LLM-based AIs into three main model types so you can make the right choice: GPT, Reasoning, and Deep Search.

1. Model number one: GPT

Used in some of the most popular platforms, like Gemini, ChatGPT, and Copilot, this model is ideal for creativity, synthesis, and open-ended questions. It’s called “Generative” because it can create new content (not just repeat information).

It’s based on pre-training, so it learns grammar, context, logic, and tone—but not real-time facts or personal experiences.

👉 Where GPT excels:

• Writing, editing, and summarizing text
• Answering simple or general questions
• Generating creative ideas and concepts (but fact-check!)
• Structuring information, like creating templates, checklists, or formatted documents

2. Model number two: Reasoning

It goes beyond recognizing patterns or generating text. It’s designed to follow step-by-step logic, allowing it to solve problems, analyze complex information, and even explain its reasoning clearly.

Reasoning models are not usually standalone platforms. They are more like an advanced mode in tools you may already know, like ChatGPT, Gemini 2.5 Pro, or Claude 4 Opus.

In these platforms, reasoning capabilities are activated or enhanced in the background, allowing the model to go beyond predicting the next likely word.

👉 Where a reasoning model excels:

• Solving multi-step problems (math, logic, or planning)
• Justifying answers with explanations
• Simulating critical thinking or alternative analysis
• Handling ambiguous or open-ended questions with structure

Model number 3: Deep Search

These tools combine the power of language models with real-time access to web information.

Unlike traditional LLMs, which rely only on the data they were trained on, Deep Search models can actively search the Internet, analyze sources, and give you updated, contextualized answers.

In short, a Deep Search platform sends your query to search engines, scans multiple pages, and uses AI to summarize and synthesize the findings—all in one response.

👉 Where Deep Search excels:

• Finds the most recent studies or news
• Evaluates multiple sources before answering
• Cites the origin of information
• Supports market or competitive research with updated data

To make it easier, we’ve prepared a simple spreadsheet to guide you, available right here in the class content.

When to use GPT, reasoning models, or deep search

Some extra tips!

#1

Every time you change topics, start a new chat. This resets the context and helps the AI stay focused, avoiding confusion from previous conversations.

#2

Typing isn’t always convenient, especially in a busy clinic. Many platforms now support voice input, allowing you to speak your question or instruction directly.

#3

Ask for a video summary: imagine watching a talk from an online medical conference but not having time to watch the full video. Simply copy the YouTube link, paste it in the chat, and request a summary.

⚙️ Prompt Engineering

After understanding the basics of LLMs, the next important topic is learning how to interact with these tools in the best way. This is where prompt engineering comes in.

First of all: what exactly is a prompt?

A prompt is the message or question you give to an AI tool to get a response. It’s how you “talk” to the AI, whether to ask for help, generate content, solve a problem, or search for information.

→ Let’s see an example: Imagine you need to clearly and fully explain to your patients what hypertension is.

Your task would be: Write a brief and clear explanation of what hypertension is so that any patient can easily understand it; Or: How would you explain hypertension to a patient without a medical background?

See? In LLM-based systems, the more precise and contextualized the information you provide, the more data the AI has to consult.

In the next lesson, you’ll see the perfect formula for building your prompt.