Perceived risks and benefits of Artificial Intelligence: A behavioral economics approach

Abstract

This paper aims to analyze individuals' perceptions of artificial intelligence (AI) and how they perceive the potential risks and benefits of AI. This study examines how information treatment effects change people's attitudes toward AI. To analyze this, an experimental survey was conducted on April 27, 2023. A total of 120 respondents were obtained through the research platform Prolific and after data cleaning, a total of 114 responses were utilized in the results. The results indicate that the following variables, income level and concern for AI, negatively impact whether individuals believe AI is harmful or helpful. Individuals also believe that AI will have a negative impact on the governmental sector. On the contrary, a positive relationship is found in self-assessed risk behavior. Furthermore, a positive relationship is observed in a variable that measures if individuals believe that AI can help humans achieve better outcomes than humans working alone. Additionally, a positive attitude toward AI is found in the following sectors: media and entertainment, financial services, financial advice, and transportation. Nevertheless, the study finds no evidence of a gender difference in the perception of AI. The study finds evidence that individuals update their beliefs when exposed to information treatments. However, the information treatment does not generate significant differences in subsequent survey questions.