The dark side of AI : A systematic literature review

Abstract: The world is currently experiencing an extraordinary explosion of data due to the advancements in digitalization, this has made the decision-making processes become increasingly complex. Modern decision-making incorporates various technologies such as AI, big data, and machine learning and they offer significant advantages in terms of speed, scalability, and granularity. However, the adoption of AI technology also has drawbacks. This paper sheds light on the potential negative outcomes that AI can produce, by systematically reviewing and synthesizing the literature. This paper provides a comprehensive overview of privacy, aversion, accountability, transparency and explainability, and bias, to give insight on the dark side of using AI in decision-making. The findings indicate that to address these issues in the dark side of AI, transparency and explainability are central. By providing explanations, ensuring privacy, and promoting transparent algorithms, we can mitigate these problems and create sustainable AI. Allowing us to embrace its potential while minimizing risks and maximizing benefits. In this paper, the aim is to contribute to the existing literature on AI-supported decision-making by presenting a broad picture and understanding of this area. Based on the findings, directions for future research in AI decision-making are presented to improve the knowledge about how to mitigate the risks of using AI in decision-making.