
EU AI Act: Europe's risk-based law for AI systems
In short: The EU AI Act regulates AI by risk level, banning some uses and setting duties for high-risk systems and general-purpose models sold or used in the EU.
The EU AI Act (Regulation (EU) 2024/1689) is Europe's risk-based law for AI systems and general-purpose AI models. It entered into force on 1 August 2024 and applies in stages. It bans some practices, sets strict requirements for high-risk systems, adds transparency duties for chatbots and generated content, and reaches providers outside the EU whose AI is placed on the market or used there.
EU AI Act in plain words
It is product-safety rules for AI: the riskier the use, the more checks before it can be sold or used in Europe.
How it works
The European Commission's AI Act page sorts AI uses into four risk levels:
- Unacceptable risk: prohibited practices such as social scoring.
- High risk: uses such as CV-sorting software for recruitment and credit scoring, which need risk assessment and mitigation, quality datasets, activity logging, human oversight, accuracy and cybersecurity.
- Transparency risk: for example, telling people they are talking to a chatbot and labelling deepfakes.
- Minimal or no risk: everything else.
Providers of general-purpose AI models carry separate duties on transparency and copyright, with extra risk work for models with systemic risk.
The timeline has moved. On the Commission's page, last updated 19 October 2026, prohibitions applied from 2 February 2025, general-purpose model duties from 2 August 2025 and the Act generally from 2 August 2026. The AI Omnibus, in force since 27 July 2026, moved high-risk duties to 2 December 2027 for listed use cases and 2 August 2028 for regulated products. Check the page before you plan.
Example
As an illustration, an Indian HR-tech company selling a CV-screening tool to European employers is likely building a high-risk system, since the Commission lists CV-sorting software for recruitment as high risk. Article 2 covers providers placing AI systems on the EU market whether they are established in the Union or in a third country, so a Bengaluru headquarters does not take the product out of scope.
Where you'll meet it
- Procurement and contract clauses from European customers
- Compliance roadmaps for HR, fintech, edtech and health products
- Model providers' general-purpose AI documentation
- Standards work translating the Act into technical requirements
What it changes for you
For product teams selling into Europe, classification of your use case is the first question. For engineers, logging, evaluation records and human oversight move from good practice to evidence. For services firms, my view is that compliance-ready delivery becomes part of the offer.
EU AI Act vs GDPR
GDPR governs personal data. The AI Act governs AI systems and models by risk. Many AI products must satisfy both.
Common mistake
My view: assuming the Act applies only to European companies. Scope follows where the system is sold or its output is used, wherever the company is registered. If your AI is used in the EU, check.
Related terms
AI governance · NIST AI RMF · ISO/IEC 42001 · Human-in-the-loop · Model card
From the archive: The AI Act delay did not stop Article 50
If your AI is sold or its output is used in the EU, the Act can apply to you wherever you are based.


