17 questions, about 8 minutes, for a first indicative estimate of where your AI system stands under the European regulation. Not legal advice.
Free indicative result · 17 questions · Email required to see your summary
144 pages in the Official Journal. Obligations that vary with the system's use, your role (provider or deployer) and your size. And a timeline in several stages.
Official application timeline
Source: Regulation (EU) 2024/1689, as amended by Regulation (EU) 2026/1744 (OJ of 24 July 2026), EUR-Lex.
Site and data hosted by Scaleway in Paris. AI drafting goes through Anthropic (United States), without your name, email or company.
Rule-based classification, obligations described article by article, and automatic checking of the articles the AI cites.
Regulation (EU) 2024/1689 as amended by the Digital Omnibus (2026/1744). The check date appears in every report.
If the report is not delivered, unreadable or incomplete, we regenerate it or refund you (see terms).
EUAICheck does not replace a lawyer. It helps you prepare the right questions before consulting one.
Each section takes into account your system's use, your role (provider or deployer), your description and your answers to the 17 questions. The result remains indicative: it depends on the accuracy of your answers.
Fictional company. French version
The category associated with your answers and the corresponding provisions of the Regulation.
Data, human oversight, logs, informing persons, roles, risk management.
Concrete, prioritised steps with an indicative effort estimate.
Ceilings set by Article 99 and factors authorities consider, without a fine estimate.
Contact
We reply within 48 hours.
EUAICheck is an independent EU AI Act self-assessment tool.
📍 Rochefort-du-Gard, Occitanie, France
Vival Solutions, EI (sole proprietorship) · SIRET 130 451 164 00011
Your data is used only to answer your message. Privacy policy.
Question 1 of 17
Your system
In one sentence, what does your AI system actually do?
No need to be technical: "It analyses CVs to shortlist candidates" is enough.
Indicative pre-assessment based on your declarations, not legal advice. Describe the system without including personal data.
System use
What is your AI system mainly used for?
What matters is how the system is used, not your company's sector. Choose the main use.
Scope of the Regulation
Is the system developed and used only for scientific research, without being sold or made available?Is the system used exclusively for military, defence or national security purposes?
Real-world testing and any availability outside research are not covered by the exclusion.If the same system is also sold or used for civilian purposes (private security, police, industry...), answer no.
Credit and insurance
What exactly does the system do?
Annex III excludes systems used only to detect financial fraud.
Essential services
Which use is it?
Who uses the system
Who uses the system?
Some obligations, such as the fundamental rights impact assessment (Article 27), depend on this.
Critical infrastructure
If the system fails or makes a mistake, what can happen?
The Digital Omnibus clarifies that a system that only assists operators or optimises performance, with no safety risk if it fails, is not a safety component.
Biometrics
What exactly does the system do?
Simple identity verification is excluded from Annex III; remote identification, categorisation and emotion recognition are listed.
Regulated product
Must your product go through a notified body (third-party conformity assessment) under its own regulation?
The AI of a regulated product is high-risk only if that product must be assessed by a third party (Article 6(1)).
Generated content
What content does the system mainly produce?
Article 50 obligations differ by type of content.
Role in the decision
What role does the system play in decisions about people?
A merely preparatory task may, under conditions, escape high risk (Article 6(3)).
Profiling
Does the system profile people (evaluate or predict their behaviour, performance, reliability...)?
With profiling, the Article 6(3) exception never applies.
Your role
What is your role regarding this AI system?
The Regulation distinguishes the provider, who develops or markets the system under its name, from the deployer, who uses it in its business. Their obligations differ.
Prohibited practices
Does your system do any of these things?
These practices are banned in the European Union since 2 February 2025 (Article 5). They are rare: serving or targeting a vulnerable public (job seekers, patients, children) is not in itself a prohibited practice. If you tick one, the next question shows you its exact conditions.
Prohibited practices: check
Does your system meet all these conditions?Does your system infer the emotions of employees, candidates or students, for a reason other than medical or safety?Does your system meet these conditions?Does your system predict offences solely from people's profiles?Does your system do this?Does your system do this?Is your system used this way, by or for law enforcement?Does your system meet all the conditions of this prohibited practice?
Article 5 bans subliminal, purposefully manipulative or deceptive techniques, or exploiting a vulnerability due to age, disability or a specific social or economic situation, where they aim at or result in materially distorting people's behaviour, leading them to a decision they would not otherwise have taken, in a way that causes or is likely to cause them significant harm. Answer "No" if you ticked this practice as a precaution and it does not match your system.Article 5 bans systems that infer a person's emotions in the workplace or in education, except for medical or safety reasons. Answer "No" if you ticked this practice as a precaution and it does not match your system.Article 5 bans evaluating or classifying people over time based on their social behaviour or personal characteristics, where the score leads to unfavourable treatment in a context unrelated to where the data was collected, or to unjustified or disproportionate treatment. Answer "No" if you ticked this practice as a precaution and it does not match your system.Article 5 bans assessing the risk that a person will commit an offence based solely on profiling or personality traits. Supporting a human assessment based on objective, verifiable facts is not covered. Answer "No" if you ticked this practice as a precaution and it does not match your system.Article 5 bans creating or expanding facial recognition databases through untargeted scraping of facial images from the internet or CCTV footage. Answer "No" if you ticked this practice as a precaution and it does not match your system.Article 5 bans categorising individuals based on their biometric data to infer their race, political opinions, trade union membership, religious or philosophical beliefs, or sex life or sexual orientation. Answer "No" if you ticked this practice as a precaution and it does not match your system.Article 5 bans real-time remote biometric identification in publicly accessible spaces for law enforcement, except in strictly listed cases with prior authorisation. Answer "No" if you ticked this practice as a precaution and it does not match your system.Answer "No" if you ticked this practice as a precaution and these conditions do not match your system.
People affected
Who is directly affected by what your AI recommends or decides?
Think about the people whose situation changes based on the outcome.
Effect on people
Can your AI's output concretely change someone's situation (hiring, refusal, grade, monitoring)?
Think about what actually happens, not what is theoretically intended.
Deployment
Is your system already deployed and used on real people?
What you already have
Could you show within 10 minutes the instructions for use and documentation supplied by the system's vendor?Could you show within 10 minutes your system's documentation: training data origin, design choices, tests?Could you show your system's documentation within 10 minutes: data origin if you built it, the vendor's instructions and documentation if you use it?
What actually exists, not what should exist.
What you already have
Before your AI's output affects a person, does a human systematically validate it?
Not the theoretical process: what actually happens day to day.
What you already have
Do you keep a record of each output your AI produces, and what it was based on?
What you already have
Are the employees affected by the system informed of it, and can they get an explanation of an outcome?Can a person affected by your AI get an explanation or contest an outcome?
What you already have
In your organisation, who is responsible for AI and compliance topics?
What you already have
Do people talking with your chatbot know it is AI?Is generated content disclosed or marked as AI-produced?Are the people concerned informed that the system analyses their biometric data or emotions?Do people who interact with your AI, or see its content, know it is AI?
What you already have
Have you formally assessed the risks your AI could cause?
Your context
What is the size of your company?
Your context
Do you have legal or compliance support?
Last step
To display your pre-assessment and email it to you.
Your data is used to display and send your pre-assessment, and to generate the report if you order it. Never sold, no solicitation. Privacy policy.
This pre-assessment is indicative. It is not legal advice, a certification or a guarantee of compliance.
Your EU AI Act pre-assessment (indicative)
Report language:
Secure payment by Stripe. Invoice sent by email. No refund, except for technical failure (terms, section 7).