EU AI Act Summary: What UK Businesses Need to Know in 2026

The EU AI Act is a comprehensive EU legal framework for artificial intelligence. It applies in phases. UK businesses can fall within scope where the Regulation's territorial conditions are met, so EU customers or activity should prompt a scope assessment rather than an assumption that every UK business is covered.

This is a plain English summary of everything you need to know.

The EU AI Act (Regulation 2024/1689) is a binding legal framework that regulates how AI systems are developed, deployed, and used. It classifies AI systems into risk tiers, from unacceptable risk (banned outright) to high risk (heavily regulated) to limited and minimal risk (lighter obligations).

The Act is designed to ensure AI used in the EU is safe, transparent, and accountable. It covers everything from AI used in hiring and healthcare to chatbots and content generation tools.

Yes, if any of the following apply to you, the EU AI Act is likely relevant:

• You have customers, users, or employees in the EU or EEA.
• Your AI system’s outputs are used inside the EU.
• You place AI systems on the EU market.

In practice, this captures the majority of UK SMEs that use AI tools like ChatGPT, automated hiring software, AI marketing tools, or AI powered customer service in any capacity that touches EU individuals.

• 2 February 2025: Prohibited AI practices banned.
• 2 August 2025: Rules for General Purpose AI models come into force.
• Article 50 contains use-case-specific transparency obligations. The applicable high-risk timetable depends on the category and current amended legislation. Article 99 provides different penalty tiers for different infringements, with special treatment for SMEs, so the €35 million or 7% headline is not a universal penalty for non-compliance.

Depending on how you use AI, your obligations may include:

• An AI Acceptable Use Policy documenting how your business uses AI.
• AI Literacy training records for staff who work with AI systems.
• Transparency disclosures when AI interacts with or makes decisions about people.
• A Human Oversight procedure for AI assisted decisions.
• A Fundamental Rights Impact Assessment, but only if you're a public body, a private provider of public services, or using AI for credit or insurance risk scoring. Most businesses using high-risk AI, including in hiring, don't fall into this category.
• Vendor AI Risk Registers documenting the third-party AI tools your business uses.

These figures apply to businesses of all sizes, but for SMEs and startups Article 99(6) caps each fine at whichever figure is lower, the fixed amount or the percentage, rather than higher. In practice this means the percentage is usually what applies to a small business.

• Up to €35 million or 7% of global turnover for prohibited AI practices.
• Up to €15 million or 3% of global turnover for non-compliance with obligations.
• Up to €7.5 million or 1% of global turnover for providing incorrect information to regulators.

The first step is understanding your risk level. Not every business has the same obligations, it depends on what AI you use, how you use it, and who it affects.

Start with our free compliance risk check at clausely.co.uk/compliance-checker to find out exactly where you stand in under two minutes.

If you already know which documentation you need, our tailored compliance packs start from £399 and are generated within the hour, ready for your review and implementation. See our packs at clausely.co.uk/pricing.