AI Act: what the EU AI regulation changes for SMEs
Risk levels, AI literacy, transparency, timeline and penalties: the essentials of the EU AI Act explained simply for SME managers.
Killian LE DIMNA Updated on 5 min read
The European regulation on artificial intelligence, better known as the AI Act (Regulation (EU) 2024/1689), entered into force on 1 August 2024. It is the world’s first legal framework dedicated specifically to AI. Many SME managers think it only concerns the tech giants. That is not true: as soon as a company uses an AI tool in its business, part of the text applies to it.
Here are the essentials, without jargon.
Provider or deployer: what is your role?
The AI Act distinguishes between several actors. The two main ones are:
- the provider, which develops an AI system and places it on the market (for example a software publisher);
- the deployer, which uses an AI system in the course of its professional activity.
The vast majority of SMEs are deployers: they use a chatbot, a CV screening tool, invoice-processing software or an automation built by a service provider. Deployers’ obligations are lighter than those of providers, but they do exist.
A risk-based approach
At the heart of the regulation is a classification of uses according to their level of risk:
- Unacceptable risk: these practices are prohibited. They include social scoring, manipulating behaviour by exploiting vulnerabilities, and emotion recognition in the workplace (except for medical or safety reasons).
- High risk: uses that are allowed but strictly regulated. They concern, for example, recruitment and employee management, access to credit, education and certain critical infrastructure.
- Limited risk: mainly transparency obligations. People must know they are talking to a chatbot, and generated or manipulated content (such as deepfakes) must be labelled.
- Minimal risk: the vast majority of uses (spam filters, writing assistance, sorting internal documents, etc.), with no specific obligation.
For an SME, the good news is that most administrative automation (invoice processing, email sorting, document summaries, lead qualification) falls under minimal or limited risk. Caution is needed, however, as soon as AI is involved in decisions about people, in particular recruitment and human resources management.
The “AI literacy” obligation, already in force
This is the point most often overlooked. Since 2 February 2025, Article 4 of the AI Act requires providers and deployers to take measures to ensure a sufficient level of AI literacy among the people who use these systems on their behalf.
In practice, this means that your staff who use AI tools must understand:
- what the tool does and does not do;
- its limits (errors, “hallucinations”, bias);
- good confidentiality practice (which data must never be entered);
- when and how human verification is needed.
The text does not impose a specific format, but it is wise to document the training you provide.
The application timeline
The AI Act applies in stages. Its timeline was amended by the so-called “Digital Omnibus on AI” regulation (Regulation (EU) 2026/1744), which entered into force at the end of July 2026 and, in particular, postponed the obligations for high-risk systems:
| Date | What applies |
|---|---|
| 1 August 2024 | The regulation enters into force |
| 2 February 2025 | Ban on unacceptable-risk practices and AI literacy obligation |
| 2 August 2025 | Obligations for providers of general-purpose AI models (those who design the large models) |
| 2 August 2026 | Transparency obligations (Article 50), for example telling people they are interacting with an AI |
| 2 December 2027 | High-risk systems listed in Annex III, including recruitment and employee management |
| 2 August 2028 | High-risk systems built into regulated products (Annex I) |
For an SME, two dates matter right now: the AI literacy obligation, which already applies, and the transparency obligations, which have applied since 2 August 2026. As the timeline may still change, check it before making any major decision.
What penalties?
The fines are high: up to €35 million or 7% of worldwide turnover for prohibited practices, and up to €15 million or 3% for most other breaches. For SMEs and start-ups, the regulation provides that the lower of the two amounts applies. The text also asks the authorities to take into account the interests and economic viability of SMEs.
The aim is therefore not to frighten, but to encourage thoughtful use.
Five habits to adopt now
- Take stock of the AI tools used in the company, including those employees use on their own initiative.
- Classify each use according to its level of risk, paying particular attention to HR uses.
- Train your teams and keep a record of this training.
- Be transparent with your customers and candidates when they interact with an AI.
- Keep a human in the loop for any important decision: AI prepares, people approve.
These habits are in line with those of the GDPR, which continues to apply in full whenever personal data is processed.
In short
The AI Act does not hold back AI adoption in SMEs: it frames it. For most administrative uses, the obligations come down to training, transparency and common sense. Companies that take these issues seriously today give themselves a real trust advantage with their customers and their teams.
Further reading
- Sovereign AI: why the hosting of your data matters
- Bringing AI into an SME: a five-step method
- Security and data: our dedicated page
This article is for information purposes only and does not constitute legal advice. For an analysis of your situation, consult a legal professional.
