Welcome to the staff secretariat AI track
Five modules in the same structure as ChatGPT for beginners: short explanation, multiple choice with feedback, and each lesson a “live exercise” you open directly in ChatGPT, Claude or Gemini. Not instruction on how AI works, actual work deliverables from your day at the cardiology department.
Which genAI tools exist, and what do they cost?
Generative AI (“genAI”) is the umbrella term for chatbots that produce text, images and code. The best known: ChatGPT (OpenAI), Claude (Anthropic), Gemini (Google), Microsoft Copilot (built on OpenAI's models, inside M365), Mistral / Le Chat (France, EU) and DeepSeek (China). All of them have a free version, that's enough to do this entire course.
| Tier | Price (indication*) | What's the difference? |
|---|---|---|
| Free | €0 | The base model, with limits on the number of messages per day and often a smaller working memory. Fine for learning and one-off tasks. |
| Paid subscription (ChatGPT Plus, Claude Pro, Gemini Advanced) | ~€20–25 per month | Access to the newest and smartest models, much higher limits, a larger working memory and extra features such as Projects and file processing. You mainly notice this with daily use. |
| Top and organisation subscriptions (e.g. ChatGPT Pro, Claude Max, Microsoft 365 Copilot at organisation level) | €100–200+ per month; on a yearly basis this can run towards €2,000 per user | Near-unlimited use of the heaviest models, priority during peak times and, with organisation licences, enterprise data protection and IT administration. You don't need this for this course. |
* Prices and packages change regularly; check the current prices with the vendor itself.
Traffic light for data (core)
Use this model before you paste anything into a public chat (ChatGPT / Claude / Gemini consumer):
| Colour | Examples | Public chat? |
|---|---|---|
| Green | Public information, self-written non-sensitive text, fictional planning, idealised practice cases without identifying details | Usually yes, check local policy |
| Amber | Internal department names, rosters without patient info, raw transcripts before cleanup, recognisable situation sketches | Only if your institution explicitly allows it; otherwise no, or anonymise first |
| Red | Patient information, BSN/SSN, EHR export, full letter or report, unique combinations (rare diagnosis + region + age) | No, only in an approved hospital environment |
When in doubt: don't paste. Better one extra step (anonymise, dummy names, fictional data) than a privacy incident.
The main vendors: who sits where, and what do you want to know first?
In this course we practise with prompts you can enter into ChatGPT or Claude. In practice you'll also encounter Gemini (built into Google Workspace), Mistral, the French / EU player, often interesting when your hospital is looking for a European supplier, and DeepSeek. The table below provides orientation, not a legal opinion. Always check legal/privacy/IT in your workplace for definitive limits.
| Name you work with | Which company? | Where is the company? | First risk topic to discuss with colleagues |
|---|---|---|---|
| ChatGPT | OpenAI | US, San Francisco. | American cloud jurisdiction; logging/training/retention varies per product and subscription; error sensitivity not to be underestimated. |
| Claude | Anthropic | US, San Francisco. | Anthropic publicly positions itself around AI ethics and safer behaviour (“AI safety”, “Constitutional AI”). No carte blanche to paste red data, your own institutional policy still rules. |
| Gemini | Google / Google DeepMind | US (Mountain View) & UK (London); global data centres, including EU. | Extra important for secretariats: Gemini in Workspace can read your Gmail / Drive / Docs if the organisation has that switched on. Check which accounts and which folders are excluded. |
| Mistral / Le Chat | Mistral AI | France, Paris, so within the EU. | Often the first answer when your organisation looks for a European supplier. EU jurisdiction and GDPR align better. Ecosystem (apps, integrations) still thinner than OpenAI/Google. |
| DeepSeek | DeepSeek | China (Hangzhou, Zhejiang). | Besides hallucinations: sovereignty / policy against non-approved cloud regions, vendor due diligence, which data layer must never travel this route. |
Question to think about: imagine your organisation had to designate one EU AI tool for public use at the secretariat. Would your department lean towards Mistral (because it's “in Europe”), or Microsoft Copilot/ChatGPT/Gemini (for the Outlook/Teams/Workspace integration)? Where would the first discussion arise in that choice?
The modules
You'll go through five modules. Each module: short explanation → multiple choice questions → live exercise in ChatGPT/Claude/Gemini → short reflection. The difference with “courses about AI” is that here you actually work in the tools every lesson.
Module 1: Meeting ChatGPT, Claude and Gemini
Open all three side by side. Ask the same question, experience the difference. First rules of thumb: which tool for which kind of work. Plus the GDPR traffic light and your own institution policy check.
Start module 1
Module 2: Mail, Teams and short messages
Short internal email (MDT/staff meeting), polite refusal, onboarding email to a new staff member in plain language, first response to a difficult email to the department head, shifting tone. Compare ChatGPT and Claude on the exact same task. First real prompt templates.
Go to module 2
Module 3: Minutes, agendas and long documents
Claude's long context (1M tokens) as the quiet superpower. Minutes from a Teams transcript, a meeting agenda from a long mail thread, a guideline summarised to one A4 for your doctor. Plus Claude Projects: your own knowledge base.
Go to module 3
Module 4: Working IN your files (Office & Workspace)
Claude directly in Word, Excel, PowerPoint (M365 add-in). Gemini in Gmail, Docs, Sheets. Microsoft Copilot briefly as reference. Plus: Claude builds a complete .pptx presentation and a working HTML tool for you. When is “working in the file” a win, and when is copy-paste faster?
Go to module 4
Module 5: GDPR, safety and your own prompt library
What do you never paste into a public chat? What do you do when a doctor drops a patient case in anyway? EU route via Mistral. Final assignment: your personal prompt library of 10–15 templates.
Open module 5After module 5: evaluation and your own prompt library
After module 5 you fill in a short evaluation (two minutes) and you get access to the assignment book: 15 extra assignments at the level of the course, to practise at your own pace with ChatGPT, Claude and Gemini. Everything stays on your device.
Go to closing and evaluation