Only a few percent of Dutch entrepreneurs have actually taken measures to bring their business into line with the legislation, according to research from the Chamber of Commerce (Chamber of Commerce). No fewer than half of the Dutch entrepreneurs indicate that they do not know the law at all.
Google’s new AI tool under fire for ‘dangerous nonsense’ about baby care: ‘This is very worrying’
Chamber of Commerce adviser Robert Bakker points to the risks that the unfamiliarity with the AI rules entails, also for small companies: “More and more entrepreneurs are using AI tools, often without knowing what the possibilities and limitations are. Results can be outdated, incorrect or discriminatory. Unsafe use, fines or fines or reputation.”
Medium-sized SMEs in particular realize that the AI Act will have consequences: 44 percent expected impact, compared to 22 percent with smaller SMEs. Yet hardly anyone is taking measures in this group either. In total, only 2 to 3 percent are active in preparations. More than half of all entrepreneurs say they have no plans.
Risk levels
The AI Act shares applications in four risk levels: minimal, limited, high and unacceptable. The higher the risk, the stricter the requirements. With high-risk applications (such as in care, education, finance or safety), strict rules regarding transparency, data quality, documentation and human supervision apply. For example, manipulative applications and social credit scores will soon be prohibited, while heavier rules apply to AI that, for example, apply applicants, assess loans or use face recognition on the street.
New chatbot in chatgpt under fire: advises ‘a lot of botox’ and losing pounds
For self -employed people and SMEs who use generative AI, such as chatgpt or Copilot, there is usually a limited risk. That means slight obligations, especially around transparency. But with everything, the entrepreneur remains ultimately responsible and the regulator can intervene in the event of incorrect or unsafe use.