EU AI Act for Manufacturing Companies

The EU AI Act applies to manufacturing companies whenever they use artificial intelligence in any form - from camera-based detection of defective products to using ChatGPT for writing production schedules or supply emails. Many Lithuanian manufacturers mistakenly believe the EU AI Act (AI Act) only concerns tech giants. In reality, Regulation (EU) 2024/1689 addresses deployers - that is, any company that uses an AI system in its operations, even if it did not create it. In manufacturing, it is especially important not to overlook AI functions embedded in equipment, as these often fall outside the scope of office employee guidelines.

Does the AI Act really apply to my manufacturing company?

Yes, if your company uses at least one AI tool. It is important to understand the difference between two roles:

• Provider - creates or sells an AI system. For example, a manufacturer of visual inspection systems.
• Deployer - uses an AI system in its operations. This is your factory if you purchase or subscribe to an AI tool.

Most Lithuanian manufacturers are deployers, not providers. The obligations for deployers are lighter, but they still exist. For more details on this, read our separate article on who the EU AI Act applies to - it is worth reviewing to accurately determine your role.

The first obligation already in effect is AI literacy. Article 4 of the AI Act has been applicable since 2 February 2025, with supervision and enforcement beginning August 2026. This article requires providers and deployers to take measures to support the development of AI literacy among staff and relevant persons, considering their knowledge, experience, training, context, and the people affected. This does not mean guaranteeing a specific or sufficient individual level of literacy.

Where is AI actually used in manufacturing?

AI in manufacturing companies appears much more widely than it seems. Here are typical examples in a small or medium-sized Lithuanian factory:

Pay attention to the last two rows. Most AI solutions in manufacturing fall under minimal or limited risk, but AI used for employee recruitment or evaluation and certain monitoring purposes can fall into high-risk categories; classification depends on the system’s intended purpose and applicable exceptions. If you want to understand how AI Act risk categories are classified, start with that article - it will help you correctly assess each tool you use.

What deadlines are relevant for manufacturing companies?

The AI Act comes into force in stages. The most important for manufacturing companies are:

1. 2 February 2025 - AI literacy obligation (Article 4) and prohibited AI practices take effect. If you monitor employees with AI tools, red lines emerge here.
2. 2 August 2026 - General application, transparency and enforcement provisions become applicable. General-purpose AI model rules began applying on 2 August 2025, subject to transitional provisions.
3. 2 December 2027 - Requirements for high-risk AI systems listed in Annex III take effect.
4. 2 August 2028 - Requirements for high-risk AI systems embedded in regulated Annex I products take effect.

In practice, this means AI literacy and documentation are relevant today, not in the future. Many factories delay this work because AI in industry seems like a "technical" rather than a "legal" issue - but the obligation arises from the fact that you use the tool, not from who created it.

What exactly does a small factory need to prepare?

Even if all your AI tools fall under the minimal risk category, AI literacy measures remain relevant, and recording them can help explain how AI use is managed. Article 4 of the AI Act does not require specific training hours, certificates, external auditors, or a dedicated AI officer. These are practical measures you can implement with our offerings.

A recommended practical package for a manufacturing company:

• AI Usage Policy - An internal document specifying which tools can be used and what data cannot be entered into them (e.g., designs, client order terms, employee personal data). Read more in AI Usage Policy in a Company.
• AI Tools Register - A list of all AI systems used: who, where, and for what purpose. It can help you explain the scope of AI use and the measures taken. See AI Tools Register.
• Compliance Record - A document recording that you have assessed risks and taken measures. This is your "paper trail" in case of an audit - see Compliance Record under Article 4.
• Employee AI Guide - A short, understandable document for shop floor workers and administrators on how to use AI safely. Example: Employee AI Guide.
• AI Literacy Training with knowledge checks so you can prove employees are trained. This training is not legally mandatory, but it is a practical way to demonstrate that measures have been taken to improve AI literacy.

The easiest way is to have a clear sequence. For this, we have prepared an AI Deployment Checklist, which lists step-by-step what to do first. The service uses templates reviewed by a compliance specialist; individual documents are generated automatically from your answers.

How does a manufacturing company differ from a service company?

In manufacturing, AI is often embedded in equipment - cameras, sensors, control systems - so employees may not even realize they are using AI. In service companies, AI is more often standalone software tools. Therefore, in a factory, AI literacy training must cover not only office staff but also production line workers and maintenance teams. If you are interested in how requirements differ, compare with how the AI Act applies to service companies.

Another risk typical in manufacturing is shadow AI. An engineer or planner might start using ChatGPT for drafting designs or emails without permission and input commercial secrets into it. Therefore, an AI usage policy in a factory is not just a formality but real protection against data leaks. Additionally, GDPR requirements must be considered: a ban on using unapproved public tools (e.g., ChatGPT) without a lawful basis, data minimisation, and appropriate safeguards is not a blanket ban on AI use but a requirement to conduct an individual assessment.

How much does this cost and how long does it take?

The work involved depends on the AI use and its risks. A small factory using ordinary office tools may be able to prepare proportionate internal guidance itself, while higher-risk uses can require specialist advice. All the documents listed above - AI usage policy, tool register, compliance record, employee AI guide, and AI literacy training with knowledge checks and certificates - are prepared automatically by nebedirbam.lt for a fixed price of €149 + VAT (€299), with no human intervention steps. You fill out a questionnaire about your company and the AI tools you use, and the documents are generated according to your case.

If you want to start with a free step, you can take free AI literacy training - this helps implement the Article 4 literacy obligation for employees and provides certificates, even if you haven’t yet purchased a documentation package. These documents and training are not legally mandatory and do not guarantee legal compliance, but they are a practical way to demonstrate that measures have been taken. The free path is for one signed-in participant; team training is included in the paid documentation package.

Quick summary for manufacturing managers

If you need to remember just a few things from this article:

• The AI Act in industry applies to deployers - that is, you, if you use at least one AI tool.
• AI literacy has been mandatory since 2 February 2025, with supervision and enforcement beginning August 2026.
• Certain AI uses in recruitment, evaluation or monitoring can be high-risk; some specific practices are prohibited. Assess the actual purpose of each system.
• A useful practical set: policy, tool register, compliance record, employee guide and training, chosen to fit your use of AI.
• Shadow AI in a factory can leak trade secrets - policy prevents this.
• GDPR does not ban AI use but requires individual assessment: lawful basis, data minimisation, approved tool, and safeguards.

For more practical articles on AI compliance, see our full list of articles.

This article’s legal information is based on official EU AI Act (Regulation (EU) 2024/1689) Article 4 and implementation timeline data provided by the EU AI Act Service Desk and the European Commission’s Digital Strategy page.