“Subjective Biased Perception on AI Knowledge and AI Aversion”

The widespread availability of Artificial Intelligence (AI) tools, particularly through wearable devices, has intensified interaction with AI and led to research questioning the human perception of AI advice. Despite of AI’s proven efficiency in various tasks, there is a notable reluctance to accept AI recommendations, known as AI aversion. This phenomenon is partly due to a mismatch between users’ understanding of AI capabilities and their expectations, which often leads to overconfidence. Efforts like the European Commission's "Ethical Guidelines on the Use of AI" endeavor to enhance AI knowledge from a young age to combat this AI aversion, but academic debate vary. This study among 110 participants showed that people’s confidence in AI knowledge influences AI aversion. The more confident an individual about their AI knowledge is, the more likely they are to reject advice from AI input in favor of human advice. The results resolve the conflicting view in the academic debate on this phenomenon and discuss the consequences of biased perceptions of AI technology when people choose between AI and human advice. We also provide evidence for governmental bodies and public to increase the effectiveness of their educational interventions on AI knowledge.