AI safety and data protection for Switzerland and the EU
Data protection and safe habits for AI use in Switzerland and the EU, for private users and small firms.

Anything you type into a cloud AI leaves your device. Treat it like a message to a third party. This page gives the frame. It is not legal advice.
The rules in short
- Switzerland. The revised Federal Act on Data Protection (FADP) governs personal data. The data protection commissioner (FDPIC) gives guidance.
- EU. The GDPR applies to people in the EU. It can also reach a Swiss firm that serves them.
- AI rules. The EU's AI Act sets duties by risk level. Its rules phase in over several years.
Private users
- Do not paste what you would not post publicly. See what never to paste.
- Look for the setting that stops your chats from training the model. Defaults differ by provider.
- Keep every API key secret and set a spending cap.
- Check facts. A hallucination can look convincing.
Small firms (KMU)
- Know what data goes in. Client data, staff data and contracts need a clear rule.
- Check the contract. Ask where data is stored, who may read it, and whether it trains models. Business plans often differ from consumer plans.
- Write a short policy. Say which tools are allowed and what may not be pasted.
- Watch for prompt injection when an AI reads mail, web pages or files.
- Consider a local model for the most sensitive work.
When in doubt
Ask your data protection adviser or the FDPIC. Do not rely on this page for a legal decision.
All pages in this section
- What not to paste into AI chatbots
Passwords, client data, IDs and secrets do not belong in a chatbot: a traffic-light checklist, provider data rules, Swiss and EU basics, and safer options.
Read next
- What not to paste into AI chatbots
Passwords, client data, IDs and secrets do not belong in a chatbot: a traffic-light checklist, provider data rules, Swiss and EU basics, and safer options.
- Hallucination
A hallucination is a fluent, confident AI answer that is false, invented or not backed by the sources it was given.
- Prompt injection
Prompt injection is an attack where text a model reads, such as a web page or file, contains instructions that make it act against its user.
- EU AI Act
The EU AI Act (Regulation 2024/1689) is the European Union's law that regulates AI systems by risk level, with duties that start in stages.