EurekaGPT
Language: English
Try it free
Features Trust Pricing
Download FAQ EU rules Contact
Try it free Sign in

AI Act

Regulation (EU) 2024/1689, as amended by Regulation (EU) 2026/1744

The EU's law on artificial intelligence. It bans a short list of practices, sets strict rules for high-risk uses such as hiring or credit scoring, and requires AI systems to say they are AI.

Plain-language summary, not legal advice. The official text prevails.

Applies in stages. Next: 2 December 2026Under review

Checked against the official sources on 28 October 2026

Explanation reviewed on 28 October 2026

Official text (EUR-Lex) Consolidated version
Under review

This explanation predates a change published on 24 July 2026. Until we update it, rely on the official text.

Official text (EUR-Lex)

When each part applies

  1. In force

  2. Banned practices, AI literacy

  3. General-purpose AI models, governance and penalties

  4. Articles 102–110

  5. Transparency and most other rules

  6. Today

  7. Machine-readable marks due for systems on the market before 2 August 2026

  8. Two more banned uses of AI

  9. General-purpose models placed on the market before 2 August 2025

  10. National sandbox

  11. High-risk uses: hiring, credit, education and others

  12. High-risk AI inside regulated products

  13. Public-authority legacy systems

What it means for you

Show the rules for

Your company

  • Do not use AI for a banned practice

    Social scoring, manipulating people, or recognising emotions at work or in education, among others. Two more bans apply from 2 December 2026.

    Art. 5

  • Help your staff use AI with care

    Take measures to support your staff's AI literacy. Since the 2026 amendment, the law does not set a level you must reach.

    Art. 4

  • Say when content you publish is AI-made

    Disclose deepfakes, and AI-written text on matters of public interest, unless a person reviewed it and takes editorial responsibility.

    Art. 50

  • Check whether your use is high-risk

    Screening job candidates, scoring credit or grading students is high-risk. From 2 December 2027 you must keep a person in control, keep the logs and inform the people affected.

    Art. 26 · Annex III

  • AI-made images, audio and video carry a mark

    Providers mark them in a machine-readable way. Whoever publishes a deepfake must say so.

    Art. 50

You as a person

  • You are told when you talk to an AI

    A chatbot or assistant must make clear it is an AI, at the latest when you first use it.

    Art. 50

  • AI-made images, audio and video carry a mark

    Providers mark them in a machine-readable way. Whoever publishes a deepfake must say so.

    Art. 50

  • Some uses of AI against you are banned

    For example manipulation that causes harm, or exploiting your age or a disability.

    Art. 5

  • You can complain

    To the market surveillance authority of your country. See who it is below.

    Art. 85

How EurekaGPT handles it

LOGICANCE S.L. provides EurekaGPT as an AI system. Mistral AI, Anthropic, Google and DeepSeek provide the general-purpose models it runs.

  • Done

    Every answer says it was generated by AI

    In the app and in every file you export.

    Terms, section 7.3 · Trust

  • In progress, due 2 December 2026

    A machine-readable mark on the files EurekaGPT creates

    So other software can tell they were made by AI.

    Terms, section 7.3

  • Done

    Banned practices are banned uses

    Our terms forbid every practice in Article 5, and decisions about a person that rest on AI output alone.

    Terms, section Annex A

  • Does not apply

    High-risk duties for EurekaGPT itself

    EurekaGPT is a general-purpose assistant. If your company uses it for a high-risk purpose, the deployer duties are yours.

    Terms, section Annex A, item 8

Who enforces it in your country

Official documents

Change log

Try EurekaGPT free

Ask a few questions free, no card needed. Plans from €14.99 a month, VAT included.

Try it free See pricing