Does the EU AI Act Apply If You Only Use AI Tools? — AI Automation article by Rui Codex

Yes, in most professional cases. Using an AI tool that a vendor built still makes your organisation an EU AI Act deployer, because the Act regulates who uses a system, not only who builds it. Some duties, such as AI literacy, already apply, while stricter high-risk duties follow on later dates.

Key takeaways

  • Using a third-party AI tool for professional work normally makes your organisation a deployer under the EU AI Act.
  • The AI literacy duty has applied since 2 February 2025 and does not depend on the risk category of the tool.
  • Most remaining obligations, including the Article 50 transparency rules, have applied since 2 August 2026, while the high-risk deployer duties for Annex III systems apply from 2 December 2027 under the amendment made by Regulation (EU) 2026/1744.
  • Your duties depend on the risk category of each use and on whether you change its purpose or behaviour.
  • A written inventory of AI tools, with their purposes and vendor documents, is the practical starting point.

Does using an AI tool make you a deployer?

Yes, in most professional settings. Under Article 3(4) of Regulation (EU) 2024/1689, a deployer is a person or organisation using an AI system under its authority, except where the use is purely personal and non-professional. Read the regulation text on EUR-Lex.

The definition does not ask who trained the model. A chatbot licence, a CV-screening feature in your HR software or a document classifier bought from a vendor all make the business that uses it the deployer. The provider carries the design, documentation and conformity duties, while the deployer carries the duties that come from putting the system to use in its own operations.

This article is general information and not legal or tax advice. Rui Codex provides technical implementation support for AI systems, and does not offer legal, compliance audit or certification services, so the reading below is an engineering view of the text that should be checked with your own advisers.

Personal versus professional use

The exemption covers purely personal activity. An employee who uses a consumer chatbot for private reading is outside the scope in that moment, while the same employee drafting client letters for the company is inside it. The test is who the use serves, not which product or subscription level is involved.

Teams often assume that a browser extension or a low-cost plan sits outside the Act. Coverage follows the use. A cheap plan used for a business decision is still a professional use.

Which deployer duties apply to the AI tools you use?

Your duties depend on the category the Act assigns to each use. Practices listed in Article 5 are prohibited, uses in the Annex III areas are high-risk and carry the detailed deployer regime in Article 26, and certain systems carry transparency duties under Article 50. Article 4 is separate: the AI literacy duty applies to providers and deployers whatever the risk level, as set out in the official text on EUR-Lex.

Most everyday office tools will not be high-risk. The classification question is therefore not about the tool category but about the decision it supports. A grammar checker and a system that ranks job applicants can run on similar underlying technology and still sit in very different categories.

Use typeQuestion to askDeployer consequence
Prohibited practiceDoes the use match a practice banned in Article 5?The use must not go ahead
High-risk (Annex III)Does the use fall in an listed area such as employment or access to essential services?Article 26 duties apply
Transparency (Article 50)Does the system interact with people or generate synthetic content?Disclosure duties may apply
Other useNone of the above appliesAI literacy and sound governance

When do the deployer dates take effect?

The Act applies in stages rather than on a single day. Article 113 of Regulation (EU) 2024/1689 sets the application dates, and Regulation (EU) 2026/1744, in force since 27 July 2026, amended parts of that timetable, so check the consolidated text on EUR-Lex before you rely on any date.

Date in the official textWhat appliesRelevance for deployers
2 February 2025Prohibited practices and AI literacyBoth apply now
2 August 2025General-purpose AI model obligations and governanceMainly providers; deployers should check vendor documentation
2 August 2026Most remaining provisions, including Article 50; Annex III high-risk rules apply from 2 December 2027Article 26 duties for Annex III systems apply from 2 December 2027
2 August 2028High-risk rules for AI in products covered by Annex IRelevant if AI is built into a regulated product

The gaps between these dates are where planning goes wrong. Teams that wait for the December 2027 date often discover that their vendor's documentation, their logging and their training records all need months of work. Starting the inventory earlier spreads that effort out.

What must high-risk deployers do under Article 26?

Deployers of high-risk systems must follow the provider's instructions, assign human oversight to competent people and monitor how the system behaves in practice. The detailed list sits in Article 26 of the regulation on EUR-Lex.

  • Use the system in line with the instructions for use supplied by the provider.
  • Assign human oversight to people with the competence, training and authority to act on it.
  • Check that input data is relevant and sufficiently representative for the intended purpose, as far as you control it.
  • Monitor operation and report serious incidents to the provider and the competent authority.
  • Keep the logs the system generates under your control for an appropriate period, which the Article sets at a minimum of six months.
  • Inform workers and their representatives before putting a high-risk system into use in the workplace.

These duties are operational. They require process design, named owners and records, which is why an engineering team often needs to be involved alongside legal counsel.

Hypothetical examples of deployer exposure

Here are two hypothetical scenarios:

  • A hypothetical recruitment team uses a licensed tool that ranks job applications. Ranking candidates falls within the employment area of Annex III, so the team would likely be a deployer of a high-risk system and would need to plan for the Article 26 duties and the worker information duty before 2 December 2027.
  • A hypothetical marketing team uses a text generator for product copy and a chatbot on its website. The generator may need only AI literacy measures. The chatbot's provider carries the Article 50(1) duty to design the system so that people know they are dealing with an AI system, and the team should confirm that the vendor has done so. Article 50 is set out on EUR-Lex.

Where does using vendor tools stop being enough?

Using a tool as the provider delivered it usually keeps you in the deployer role. Changing the tool can move you into the provider role, which carries a heavier set of duties.

Under Article 25, a deployer can become a provider of a high-risk system if it places its name or trademark on it, makes a substantial modification, or changes its intended purpose so that the system becomes high-risk. Fine-tuning a model for a new decision, or building an internal screening tool on a foundation model, can cross that line. The same Article 25 text on EUR-Lex sets out the conditions.

Vendor-based approaches also fail in predictable ways:

  • The vendor's documentation does not describe the intended purpose clearly enough to classify your use.
  • Staff adopt AI tools that IT has never recorded, so no inventory exists.
  • Logs are kept only by the vendor and cannot be retrieved for your own oversight.
Tip: Record the intended purpose of every AI tool in writing. The purpose decides the risk category, and a change of purpose is an easy place for the deployer role to shift without anyone noticing.

What should an operations or IT lead do first?

Start with an inventory, then classify each use, then compare the findings with vendor documents and the dates above. Each step depends on the one before it, so skipping the inventory usually means the classification is guesswork.

  1. List every AI feature in use, including features embedded in software you already run.
  2. Record the purpose, the users and the data each use touches.
  3. Classify each use against Article 5, Annex III and Article 50.
  4. Collect the provider's instructions for use and its compliance documentation.
  5. Set AI literacy measures that match each role's technical knowledge and context of use.
  6. Review the inventory whenever a vendor adds a feature or you change a purpose.

Technical support is often the bottleneck. Rui Codex is led by a founder with 200+ projects delivered, and the company offers implementation work such as AI consulting and AI automation, including logging, oversight workflows and integration of AI tools with existing systems. For the data protection side of these projects, the security and GDPR engineering articles cover the technical controls in more detail.

Frequently Asked Questions

Does the EU AI Act apply to a company that only buys AI software?

Yes, in most professional cases. The Act covers deployers, meaning organisations that use an AI system under their authority outside purely personal use, so buying a vendor's tool is normally enough. The vendor keeps its provider duties, and your organisation has its own deployer duties on top of them.

Is AI literacy required for every employee who uses AI tools?

Article 4 requires providers and deployers to take measures ensuring a sufficient level of AI literacy among staff and others who operate AI systems for them. The measures should reflect each person's technical knowledge, experience, education and the context in which the system is used. The duty has applied since 2 February 2025 and does not depend on the risk category of the tool.

What happens if a deployer changes the purpose of an AI tool?

The deployer's role can change. Under Article 25, a deployer that makes a substantial modification or changes the intended purpose so that a system becomes high-risk can be treated as a provider. Review your inventory whenever a purpose changes, and check the current wording on EUR-Lex.

Which date matters most for deployers of high-risk AI tools?

For Annex III systems, the Article 26 deployer duties apply from 2 December 2027, as amended by Regulation (EU) 2026/1744. Systems that are part of products covered by Annex I follow from 2 August 2028. Confirm the current dates in the consolidated text on EUR-Lex before you plan the work.

Conclusion

Using an AI tool at work brings your organisation into the deployer role in most cases, and some of the duties attached to that role already apply. The practical work starts with an inventory of tools, a written purpose for each use and a check of vendor documentation against the dates in the regulation. Most of that work is documentation and process design, and it is easier to complete before a deadline than during one. If you want technical help mapping your AI tools and the integrations around them, request a project consultation at https://ruicodex.com/Contact.

Sources

Tags: EU AI Act AI deployers AI compliance AI literacy high-risk AI AI automation Belgium AI governance Article 26

Need Help Implementing This?

Our team can help you put these insights into practice. From AI automation to custom software development, we build solutions that deliver real results.

Send a project request