What Is the EU AI Act? A Beginner's Guide to the Rules
The EU AI Act is a European law that sets risk-based rules for artificial intelligence systems sold or used in the European Union. Officially called Regulation (EU) 2024/1689, it is described by the European Commission as the first comprehensive legal framework on AI. It entered into force in August 2024 and applies in phases, with stricter duties attached to higher-risk uses.
The short definition
The AI Act does not ban AI. It sorts AI uses by the risk they pose to people's safety and fundamental rights, then assigns obligations that rise with the level of risk. Most everyday AI falls into the lowest category and carries few or no duties, while a small set of uses is prohibited.
The four risk categories
| Risk level | What it means | Typical examples |
|---|---|---|
| Unacceptable | Uses not allowed on the EU market | Certain manipulative or harmful practices |
| High | Strict requirements and oversight | Some uses in health, hiring, or critical infrastructure |
| Transparency | Users must know they are interacting with AI | Chatbots, some AI-generated content |
| Minimal | No specific obligations | Spam filters, many recommendation systems |
Unacceptable risk
A narrow set of practices is prohibited because they are considered to threaten people's rights or safety. These systems cannot be placed on the EU market.
High risk
High-risk systems face the heaviest requirements, such as risk management, data quality, documentation, human oversight, and accuracy standards. The exact duties depend on the use case.
Transparency risk
Systems such as chatbots should make it clear that a person is interacting with AI. Some AI-generated content also needs labeling so people are not misled.
Minimal risk
Most AI systems fall here and face no mandatory obligations under the Act. Organizations may still follow voluntary codes of conduct.
General-purpose AI models
The rules also cover general-purpose AI (GPAI) models, the powerful models that can power many different applications. Providers of these models have transparency and risk-management duties, with extra requirements for models considered to have systemic impact.
Who the rules apply to
The Act names several roles rather than applying to one type of company.
- Providers that develop an AI system or model.
- Deployers that use an AI system.
- Importers and distributors that bring or place systems on the market.
Obligations differ by role, so a business building a tool and a business using one may have different duties.
How enforcement works
Enforcement is shared between EU Member States and the European Commission, a model described as a hybrid enforcement system. Bodies such as the European AI Board, a panel of independent experts, and an advisory forum support the process. This means oversight happens at both national and European levels rather than through one single regulator.
How the rules roll out
The Act entered into force in 2024 but its obligations apply in stages rather than all at once. Prohibitions take effect earlier, while high-risk and other requirements phase in over time. If you need exact dates for a specific obligation, check the official European Commission and EUR-Lex sources, because the timing differs by category.
Why it matters beyond Europe
- Companies that want access to the EU market may need to meet the rules even if they are based elsewhere.
- The risk-based approach is being watched by other countries considering similar laws.
- It gives organizations a shared reference for trustworthy AI practices.
Frequently asked questions
Is the EU AI Act a ban on artificial intelligence? No. Most AI is in the minimal-risk category with no mandatory obligations. Only a small set of practices is prohibited.
When did the AI Act enter into force? It entered into force in August 2024, with obligations applying in phases rather than all at once.
Does it apply only to European companies? It applies to systems placed on the EU market or used in the EU, so organizations outside Europe can still be affected.
What is a general-purpose AI model? It is a model that can be used for many different applications, and providers of these models have specific transparency and risk duties.
Where can I find the exact legal requirements? The official EUR-Lex text and European Commission guidance are the primary sources and should be checked for precise obligations and dates.
Key takeaways
- The EU AI Act (Regulation (EU) 2024/1689) is a risk-based framework for AI in the European Union.
- It groups uses into unacceptable, high, transparency, and minimal risk, with duties rising with risk.
- It also covers general-purpose AI models and several roles, including providers and deployers.
- Enforcement is shared between Member States and the Commission, and rules apply in phases.
Read next: what AI prompt injection is, how AI privacy protection works, and what AI-generated content is.
Sources
- https://digital-strategy.ec.europa.eu/en/policies/regulatory-framework-ai
- https://eur-lex.europa.eu/EN/legal-content/summary/rules-for-trustworthy-artificial-intelligence-in-the-eu.html
- http://ec.europa.eu/commission/presscorner/detail/en/ip_24_4123
- https://www.europarl.europa.eu/RegData/etudes/ATAG/2026/785670/EPRS_ATA(2026)785670_EN.pdf
Learn by doing
Read a guide, then practice with bite-sized AI lessons you can take offline.
Start learning free