The European Artificial Intelligence Act (AI Act) has been in force since 2024 and will be fully implemented from 2026. Yet, according to new research by the Netherlands Chamber of Commerce (KVK), most Dutch entrepreneurs remain largely unprepared. Half of all business owners are unaware of the legislation, only 7% are well-informed, and just a few percent have taken concrete steps to comply.
Limited awareness, even among larger SMEs
Medium-sized SMEs (with 10 or more employees) are the most aware of the upcoming changes: 44% expect the AI Act to affect their business operations, compared to 22% of smaller SMEs. However, even in this group, few have taken any action. In total, only about 2–3% of entrepreneurs are actively preparing, and more than half say they have no plans to do so.
A gap between AI use and understanding
While entrepreneurs are increasingly discovering AI tools in practice, their knowledge of the related legal requirements lags behind. Three in ten are already using applications such as ChatGPT or Microsoft Copilot — a figure that rises to 46% in the business services sector. One in four entrepreneurs expect generative AI to become important for their company, and among medium-sized businesses, that figure is as high as 50%. Entrepreneurs who are familiar with the AI Act are also more likely to recognise AI’s strategic importance.
Uncertainty and confusion remain high
Four in ten entrepreneurs are unsure whether the AI Act even applies to their business. Many, particularly self-employed professionals (zzp’ers), say they do not know where to start. This uncertainty reflects the overall lack of awareness and clarity about how the legislation will impact different sectors.
What the AI Act requires
The AI Act classifies applications into four risk levels: minimal, limited, high, and unacceptable. The higher the risk, the stricter the requirements. High-risk systems — for example in healthcare, education, finance, or security — must meet strict standards for transparency, data quality, documentation, and human oversight.
Importantly, not only developers but also users of AI systems are responsible for correct and lawful use. They must determine whether a system falls under the law, make clear agreements with suppliers, be transparent with customers (for instance, that a chatbot is not human), and ensure staff receive proper training. Some requirements will take effect gradually from 2025, with full enforcement in 2026.
AI tools like ChatGPT and Copilot
For most self-employed professionals and SMEs using generative AI tools, the risks are considered limited. This means lighter obligations, primarily around transparency. However, entrepreneurs remain fully responsible for how AI is used — and regulators can intervene if misuse or unsafe practices occur.
Risks for all businesses, big or small
Robert Bakker, adviser at KVK, explains: “More and more entrepreneurs are using AI tools, often without a clear understanding of their capabilities or limitations. The results can be outdated, inaccurate, or even discriminatory. It doesn’t surprise me that most business owners don’t yet know much about the AI Act, but this lack of awareness does carry risks — even for small companies. Unsafe use can lead to legal problems, fines, or reputational damage. KVK informs and advises entrepreneurs on these topics.”
The KVK will further explore the topic during its Entrepreneur Days (KVK Ondernemersdagen) on 14 and 15 November, with a series of inspiration sessions about the responsible use of AI.
About the research
The findings come from the KVK Omnibus Survey, June 2025 edition. A total of 651 entrepreneurs — including both self-employed professionals and SMEs — completed an online questionnaire on various business topics.
Text: KVK
Translation & Adaptation: The Hague Online