Inquisitor Labs

Check if the EU AI Act Applies to You

Most developers and development companies underestimate how aggressively the EU AI Act reaches across borders. The definitions are broad, the scope is wide, and regulators assume your system is covered unless you can prove otherwise. It does not matter where you are based, what you intended, or how small your operation is—if your system touches the EU in any meaningful way, you are treated as within the framework.

This page gives you a clear, practical way to understand how the Act sees you. Not how you see yourself. Not how you would prefer to be classified. How regulators will classify you when your system crosses their threshold.

The questions below map directly to the kinds of triggers regulators use. Answer them based on operational reality and you will know whether you are facing a minor administrative task or a full compliance burden you cannot ignore.

The Quick Checker

The EU assesses AI systems based on impact, jurisdiction, and accessibility. Their criteria are expansive and applied without regard to your size, location, or business model. If your system interacts with the EU in any operational way, it is treated as within their authority.

These are the kinds of triggers used to determine liability under the EU AI Act. Answer them based on how your system actually operates.

  1. Use in the EU: Is anyone in the EU using your system, directly or indirectly?
  2. Outputs into the EU: Do your outputs reach people or organisations in the EU, even through another product?
  3. EU deployment: Is your system deployed by an EU‑based team, client, partner, or contractor?
  4. Decisions about individuals: Does your system influence decisions about individuals in the EU?
  5. EU‑related data: Does your model process data relating to EU citizens or residents?
  6. EU‑available products: Is your system integrated into a product or service available in the EU market?
  7. Accessible endpoints: Do you provide an API, model endpoint, or tool that EU users can access?
  8. Significant effects: Does your system support or automate decisions with legal or significant effects?
  9. Scoring and assessment: Does your system classify, score, predict, or assess individuals?
  10. Influence and impersonation: Does your system generate content that could mislead, impersonate, or influence behaviour?

What Your Answers Mean

If you answered YES to anything in 1–7

You are liable under the EU AI Act. Your system falls within the EU’s regulatory perimeter. This is not optional and does not depend on your intention, size, or jurisdiction.

If you answered YES to anything in 8–10

You may be liable for additional risk‑based obligations, including documentation, testing, and behavioural evidence. The more your system affects individuals and decisions, the heavier the expected controls.

If you answered NO to everything

You are not currently liable under the EU AI Act. However, liability begins the moment your system, its outputs, or its endpoints enter the EU ecosystem in any form. Treat this as a point‑in‑time assessment, not a permanent exemption.

What Happens Next

If this checker indicates liability, regulators will expect you to determine and document at least the following:

None of this is optional once you are liable. The only question is whether you define this structure yourself, or have it defined for you under pressure.

Get Compliant Now

If this page has made it clear that you are liable under the EU AI Act, the next step is not to read more theory. You need a structured, practical route to compliance that matches how your system actually behaves in the real world.

The Get Compliant Now page outlines a concrete path: focused on testing, evaluation, and behavioural evidence that stands up when regulators ask difficult questions.

You do not have to deal with this alone. But you do have to deal with it.

Further EU AI Act Resources