Supporting Small and Medium Enterprises in AI Development: The Role of EU AI Regulatory Sandboxes

Abstract

Small and medium-sized enterprises (SMEs) make up a large portion of the global economy and their contribution to innovation, employment and economic growth is significant. Artificial intelligence (AI) is rapidly spreading, and it has created opportunities for productivity and innovation. The limited resources of SMEs often hinder their ability to develop and adopt new AI innovations which could be a significant competitive disadvantage. The newly developed AI legislation by the European Union, the AI Act, introduces a framework for AI regulatory sandboxes which are designed to help companies, especially SMEs and start-ups, to safely develop and test new AI technologies without regulatory barriers. Regulatory sandboxes have been used the fields of financial technology and privacy, and they have been shown to have several beneficial aspects for companies. The purpose of this thesis is to research whether the AI regulatory sandboxes presented in the EU AI Act could aid SMEs in AI development and innovation. In this thesis the challenges of European competitiveness are presented followed by digital transformation in SMEs. After that, the basis for the EU AI Act and the Act itself are presented followed by current AI regulations. Next the thesis goes more deeply into the existing regulatory sandboxes followed by the empirical part of this thesis. It consists of analysing two questionnaires which aim to map out the attitudes of SMEs towards AI development, AI regulation and the AI regulatory sandboxes. European competitiveness faces challenges, and Europe is falling behind in the tech market while the United States and China are making advancements. SMEs, taking up a large share of the European economy, are important to consider when trying to improve the competitiveness of Europe. Innovation adoption, which would accelerate growth, is often difficult for SMEs as they do not possess the needed resources. Due to the risks that artificial intelligence poses, EU has taken an initiative to develop a comprehensive AI regulation, the AI Act. The AI Act applies to the EU market and actors operating in the market using AI must comply with it. The AI Act includes a framework for the regulatory sandbox and taking into consideration the limited resources of SMEs EU has provided them with a free access to the sandbox. Regulatory sandboxes from the fields of financial technology and privacy have been shown to have benefits for companies such as facilitated market entry and they have helped especially smaller companies to navigate the complex regulatory environment. However, they have possible risks for example for competition and consumers if not designed and managed properly. Results from the empirical part of this thesis show that the regulatory sandbox for artificial intelligence can offer help to SMEs in navigating the regulatory environment and thus develop more trustworthy AI solutions as well as provide regulatory guidance for SMEs. The sandbox can offer a safe environment to develop test and validate AI innovations as well as allows for better communication and cooperation between different actors such as companies and regulators.