Your EU AI Act resource hub
Make sense of the EU AI Act, fast.
Plain-English answers, free tools, and updates you can trust, for any company that builds or uses AI. We read the regulation so you don't have to.
Find your way through the EU AI Act.
The phased deadlines
2 August 2026 is the date most companies are counting down to: most high-risk and transparency obligations begin to apply.
Prohibited practices + AI literacy.
The Article 5 bans (e.g. social scoring, manipulative AI) and the Article 4 duty to ensure staff AI literacy already apply.
GPAI rules, governance & penalties.
General-purpose AI model rules, the AI Office, and the penalty regime are now in force.
Most high-risk + transparency obligations.
The big one: most Annex III high-risk obligations and the Article 50 transparency duties begin to apply.
AI as a safety component of regulated products.
High-risk AI embedded in Annex I products (e.g. machinery, medical devices) and pre-Aug-2025 GPAI models must be in compliance.
⚠️ A proposed Digital Omnibus on AI would push the Annex III high-risk date from 2 August 2026 to 2 December 2027. As of 9 June 2026 this is PROPOSED, not yet law, so the binding dates above still stand. We'll tell you the moment that changes.
What applies to me? →Just been told your AI system might be “high-risk”? Start here.
Your legal team, a customer, or a board member just flagged the EU AI Act, and now you're trying to work out whether the tool you built or bought is “prohibited,” “high-risk,” or barely touched by the rules at all. The deadlines, the jargon and the talk of seven-figure fines make it feel bigger than it is.
Take a breath. The AI Act sorts AI into four risk tiers, and most everyday systems land in the lightest ones with few or no new duties. The work is figuring out which tier your system falls in, and whether you're the “provider” who builds it or the “deployer” who uses it. We'll help you do exactly that, in plain English. No jargon, no sales pitch.
Not sure if the AI Act even applies to you? Find your risk tier in two minutes →
Start here
Start where you are
Four routes through the regulation, depending on what you need right now.
Free tools
Free tools, no email wall
Use them on the page. We'll only ask for your email if you want your result or a PDF sent to you.
Why use this hub
Why use this hub
No upsell
We're not selling you compliance software, and there's no demo to book. That means no pressure to "request a quote" at the bottom of every answer. We just explain the rules.
Always current
The AI Act is moving fast: a proposed Digital Omnibus that could shift the high-risk deadline, the GPAI Code of Practice, and harmonised standards still landing. Every page carries the sources we used and the date we last checked them, and we update when things change.
Plain English, with free tools
A risk-tier classifier, an Annex III high-risk lookup, a glossary and a deadline tracker, written for people without a legal team. Terms are explained the first time we use them, then linked to the glossary.
The AI Act Brief
We watch Brussels so you don't.
The EU AI Act is still in motion. One email, plain English, tells you what changed, what it means for you, and what to do about it. So you can stop refreshing EUR-Lex and get back to your actual job.
- What the AI Office and the GPAI Code of Practice mean for you, in plain terms.
- Breaking-change alerts the moment something material lands: a Digital Omnibus deadline shift, a new harmonised standard, or a fresh prohibition.
- Plain-English summaries with a link to the official source, every time.
The AI Act Brief
Enter your email to subscribe
Free, and you can unsubscribe in one click. No spam, no selling your address.
No spam. Unsubscribe anytime.
By the numbers
The EU AI Act in a few numbers
The figures worth keeping in your head. Each one is set in law or in the official guidance we link from the pillar page.
Risk tiers the Act sorts AI into: prohibited, high-risk, limited (transparency) risk and minimal risk.
The maximum fine for breaching the Article 5 prohibited-practice bans: €35 million or 7% of worldwide annual turnover, whichever is higher.
When most high-risk (Annex III) and Article 50 transparency obligations begin to apply, unless the proposed Digital Omnibus postpones it.
Annex III high-risk areas, from biometrics and employment to credit scoring and law enforcement.
Penalty tiers run €35m / 7% for prohibited practices, €15m / 3% for other breaches and €7.5m / 1% for supplying misleading information. For SMEs and start-ups the lower of the two figures applies.
AI Act FAQ
Common EU AI Act questions
What is the EU AI Act?
The EU AI Act is Regulation (EU) 2024/1689, the world's first comprehensive law on artificial intelligence. It takes a risk-based approach, sorting AI into four tiers, prohibited, high-risk, limited (transparency) risk and minimal risk, with obligations that scale with the risk. General-purpose AI models are regulated on a separate track. It entered into force on 1 August 2024 and applies in phases through 2027.
Is my AI system in scope, and what risk tier am I?
Most AI is in scope, but most everyday systems land in the lighter tiers with few or no new duties. A handful of practices are banned outright (Article 5); a defined set of use cases, the eight Annex III areas plus AI built into regulated products, are high-risk and carry the strictest obligations; chatbots and AI-generated content carry transparency duties; everything else is minimal-risk. The free risk-tier classifier gives you a plain-English answer.
What are the EU AI Act deadlines?
The Article 5 prohibitions and the Article 4 AI literacy duty have applied since 2 February 2025. GPAI model rules, governance and penalties have applied since 2 August 2025. Most high-risk (Annex III) and transparency obligations apply from 2 August 2026, and AI embedded in regulated products from 2 August 2027. A proposed Digital Omnibus would push the Annex III high-risk date to 2 December 2027, but as of June 2026 that is not yet law. See the full timeline tracker.
Who has to comply, and does the AI Act apply to non-EU or US companies?
Duties fall mainly on providers (who develop and place an AI system or GPAI model on the market) and deployers (who use one under their own authority in a professional capacity). Importers, distributors and product manufacturers also have roles. Crucially, the Act has extraterritorial reach: a non-EU or US company is in scope if it places AI on the EU market or if the output of its AI is used in the EU. The obligations guide breaks down provider vs deployer duties.
What are the penalties under the EU AI Act?
Fines are tiered. Breaching the Article 5 prohibited practices can cost up to €35 million or 7% of worldwide annual turnover, whichever is higher. Most other breaches are capped at €15 million or 3%, and supplying misleading information at €7.5 million or 1%. For SMEs and start-ups the lower of the two figures applies. The penalty regime has been in force since 2 August 2025.
Stay ahead of the next AI Act change.
Free, plain-English updates. We watch Brussels so you don't.