MMOOZZGG.com

Glossary

EU AI Act explained: what the AI Act is

In short

The EU AI Act is Regulation (EU) 2024/1689 on artificial intelligence. It sorts AI systems by level of risk and sets requirements for each level.

Security and law · EU AI Act

MMOOZZGG.com editorsChecked: 1 min read

In depthEU AI law 2026: what is already in force

For a child

A swimming pool has rules: only those who can swim may go into the deep end, and anyone may splash in the kids’ pool. The more dangerous the place, the stricter the rules. Europe made similar rules for smart programs: simple helpers need almost nothing, and those that decide important things for people need many checks.

In plain words

This is a European Union law on AI. It does not ban neural networks as a whole; it looks at where and why they are used. Some practices are banned outright, such as social scoring of citizens or recognizing employees’ emotions at work. Systems for hiring, credit scoring, education and critical infrastructure count as high-risk: they are tested, documented and kept under human oversight. A chatbot must say that it is not a person, and generated content gets a label.

For business

The regulation covers more than European companies: it applies if an AI system enters the EU market or its output is used in the EU. The bans apply from 2 February 2025, the rules for general-purpose models from 2 August 2025, and the transparency requirements from August 2026. The requirements for high-risk systems moved to 2 December 2027, and for systems inside regulated products to 2 August 2028. The fine for a banned practice reaches 35 million euros or 7% of worldwide turnover. If you have customers in the EU, check whether your chatbot tells them that it is AI.

For a specialist

Regulation (EU) 2024/1689 entered into force on 1 August 2024 and sorts systems into four levels: unacceptable risk (Article 5), high risk, transparency risk and minimal risk. Article 5 bans, among other things, untargeted scraping of facial images for recognition databases and emotion inference at work and in schools. Before it reaches the market, a high-risk system gets a risk assessment and mitigation, quality data, activity logs, documentation and human oversight. Under Article 53, providers of general-purpose AI models respect rights holders’ opt-outs from text and data mining and publish a summary of training data. The AI Omnibus package entered into force on 27 July 2026 and moved the dates for high-risk systems. Under Article 99, the fine for breaching other obligations reaches 15 million euros or 3% of turnover.

Example

A Russian company sells online courses to customers in Germany and answers them through a chatbot. To meet the European transparency rules, the bot says in its first message that it is an AI assistant and passes the chat to a staff member on request.

Related terms

Questions and answers

Does the EU AI Act apply to Russian companies? Yes, if your AI system enters the EU market or its output is used in the EU. The regulation does not apply directly to companies that work only in Russia.
When did the EU AI Act enter into force? The regulation entered into force on 1 August 2024 and applies in stages: the bans from February 2025, the transparency rules from August 2026. After a delay, the requirements for high-risk systems start in December 2027.

Sources

Not sure whether your company needs this?

AI solutions map
Contact us