Responsible use of AI

Four questions to consider before using AI in anything that counts:

1. Is it allowed here? Workplace rules take precedence over everything else. Ask if it is not stated.
2. What am I inputting? Personal data (names, national IDs, health), customer data, and trade secrets should not be entered into external services. What you write may be stored and sometimes used for training.
3. Who owns the result? AI-generated text and images have unclear copyright protection, and the underlying material may contain others' works. Imitating a named artist's style is debated; copying protected material is not.
4. How do I document it? One line is enough: what AI was used for and what you did yourself.

The framework in brief:

Rule	What it says	Practical implication
GDPR	Personal data may only be processed with a legal basis and for a specified purpose	Do not enter names, addresses, or health details of customers or colleagues
Copyright	Works are protected automatically; copying requires permission	Do not paste in entire protected texts; cite sources
EU AI Act	Classifies uses by risk; high-risk uses (recruitment, credit) require documentation and human oversight	An AI system filtering CVs must not decide alone
Workplace rules	Local application	Determines what is allowed in the specific task

Documentation template: «I used an AI assistant to get three suggestions for the introduction and for proofreading. All text was written by me, and facts were verified against [source].»