The use of artificial intelligence in companies’ decision-making

Advanced technologies such as intelligent algorithms and artificial intelligence increasingly provide companies with powerful new methods for decision-making. Research on changes in decision-making is interesting and important. Data-driven decisions are not new for companies and there is a high likelihood that increasingly routine business decisions will be made by machines. A central feature of artificial intelligence is that it learns from data. Knowledge can be incorporated in no other way. That means that results that the artificial intelligence produces will reflect any inaccuracies in that data. Also, many artificial intelligence devices are so complex that it cannot always be explained how they arrive at answers. The use of artificial intelligence in companies’ decision-making is growing and this thesis therefore examines the intersection of artificial intelligence and company law, and analyzes how they fit together. The study aims to find the boundaries of the use of artificial intelligence in companies’ decision-making, namely to what extent company directors should be allowed to rely on artificial intelligence. This is done by analyzing the significance of information and its importance for decision-making. Moreover, the research examines how it can be ensured that companies’ use of artificial intelligence is fair, accountable, and transparent. This thesis focuses on the European Union and Finnish law framework, following the traditional legal dogmatic approach. These research findings indicate that information technology has made business more effective, but currently there still exist problems when artificial intelligence is involved in companies’ decision-making processes. It is also unknown how modern technology can handle situations where decision-making requires creative reasoning and an understanding of the bigger picture, as creativity and innovation are widely regarded as specifically human qualities. Legislation and guidance should define artificial intelligence’s liability issues, licensing practices, and transparency requirements so that its usability could become as inclusive as possible