FAQS: EU Artificial Intelligence Act (EU AI Act)

What is the AI Act?

The EU Artificial Intelligence Act is the European Union’s new law designed to make sure AI systems are used safely, responsibly and in line with European values. It sets out how AI can be developed, placed on the market and used – both within the EU and by companies outside the EU that want to offer AI systems to EU customers.

At its core, the AI Act takes a risk-based approach. The higher the potential risk an AI system poses to people’s rights or safety, the stricter the requirements will be. Some uses of AI that are seen as a clear threat to human rights will be banned altogether.

As the first comprehensive AI law in the world, the AI Act is likely to set a global standard for how artificial intelligence is regulated and trusted.

What products are affected by the AI Act?

The AI Act doesn’t just apply to one type of product – it covers any product or service that includes an AI system and is placed on the EU market. This includes standalone software and AI that’s built into other products, such as:

• Smart home devices and connected appliances
• Autonomous or driver-assistance systems in vehicles
• AI-powered medical devices and diagnostic tools
• Industrial machinery and robotics
• Consumer electronics with speech or image recognition
• AI tools used in recruitment, education or public services

If a product uses AI to make decisions, predictions, or automate processes, it may fall under the scope of the AI Act.

For most cases, the AI system is regulated in the same way as the product it forms part of – for example, under existing CE marking rules for machinery, medical devices or toys. However, the AI Act adds an extra layer of transparency and risk management requirements, especially for high-risk AI systems.

When does the AI Act come into force?

The EU Artificial Intelligence Act entered into force on 1 August 2024, with its requirements applying in stages.

Prohibited AI practices and certain AI literacy requirements have applied since 2 February 2025.

Further requirements, including rules relating to general-purpose AI models, began applying from 2 August 2025.

The AI Act became generally applicable on 2 August 2026, although some requirements have later application dates.

Following changes to the implementation timetable, requirements for certain high-risk AI systems are being introduced later. High-risk systems covered by Annex III are due to comply from 2 December 2027, while requirements for high-risk AI systems that are safety components of, or are themselves, products covered by specified EU product safety legislation are due to apply from 2 August 2028.

Businesses should therefore identify how their AI system is classified under the Act and which implementation date applies to i

What about products that are already on the market?

If your product is already on the EU market before the AI Act starts to apply, it won’t be forced to stop selling immediately. However, any new AI systems placed on the market or put into service after the relevant dates will need to meet the AI Act’s requirements.

That means existing products can continue to be made available for a limited time, but updates, new models, or significant software changes could bring them under the new rules.

If an AI system is substantially modified after the law takes effect, it will likely be treated as a new system, and full compliance will be required.

Manufacturers should start reviewing their products now to identify any AI functionality and prepare the necessary technical documentation, risk assessments, and transparency measures ahead of the transition period.

What about products that are already on the market?

If my product is already on the market, will I need to change it to meet the regulation?

Not necessarily – but it depends on how your product uses AI and whether it will continue to be sold or updated after the AI Act starts to apply.

If your AI-enabled product is already on the EU market before the relevant deadlines, it can generally stay there without immediate changes. However, once the AI Act becomes applicable, any new versions, major software updates, or re-releases will need to comply with the new rules.

If your AI system is classed as high-risk, you may also need to review your technical documentation, risk management processes, and transparency information to make sure they meet the new requirements.

In short – you won’t have to recall existing stock, but you should prepare now to ensure future models and updates align with the AI Act’s standards.

What documentation will I need to prepare and make available for the surveillance authorities?

Under the AI Act, companies placing AI systems on the EU market – particularly high-risk AI systems – must prepare and maintain detailed documentation to show compliance. This documentation should be ready to present to market surveillance authorities on request.
In most cases, you’ll need to have:

Technical documentation describing how the AI system is designed and operates, including its purpose, key components, and algorithms used.

Risk management records, showing how potential harms have been identified, assessed, and mitigated throughout the product’s lifecycle.

Data and training documentation, explaining the quality, origin, and representativeness of the data used to train the AI model.

Post-market monitoring plans, outlining how you will track performance, incidents, and feedback once the product is on the market.

Conformity assessment reports (for high-risk systems), including any test results and certificates from notified bodies, where applicable.

Instructions and transparency information provided to users, showing how to operate the system safely and understand its limitations.

This documentation must be kept up to date and retained for at least 10 years after the AI system is placed on the market.

What role does an Authorised Representative play in the process?

An European Authorised Representative (AR) acts as your official point of contact within the EU for regulatory and market surveillance authorities. If your business is based outside the EU but you place AI systems or AI-enabled products on the EU market, appointing an Authorised Representative is a legal requirement under the AI Act.

The Authorised Representative’s role is to:

• Hold and provide access to your technical documentation and declarations of conformity when requested by authorities.

• Cooperate with market surveillance authorities to respond to questions, provide information, or support inspections.

• Ensure communication between your company and EU regulators is reliable and efficient.

• Verify that your conformity assessment and declarations have been completed correctly before the product is placed on the market.

In short, your Authorised Representative helps make sure your AI system can be legally placed on the EU market – and that you remain compliant throughout its lifecycle.