Racial Disparity In Perceived Effectiveness And Equity Of Artificial Intelligence

This study introduces perceived effectiveness and equity as key dimensions of public perception of artificial intelligence (AI) and examines racial disparities in these perceptions. Using the American Trends Panel survey by Pew Research Center, this study examines how White, Black, and Asian respondents perceive effectiveness and equity in AI's application overall and in different fields. Findings show that both Black and White respondents have a relatively lower level of overall perceived effectiveness of AI, while Asian respondents have a higher level of effectiveness perception. For specific fields of AI application, Black respondents have a lower level of perceived AI’s effectiveness in detecting cancer and producing crop than the other groups, while White respondents have a lower level of perceived effectiveness of AI’s application in mental healthcare, detecting protein structure, and writing news, suggesting Black is more cautious about AI’s application in fields that are directly related to resources and personal interests, while White is less optimistic about AI’s applications in fields that involve more personal components or personalization. In terms of perceived equity, Black respondents report a lower level of perceived equity overall, as well on healthcare and hiring, which goes against previous expectations that AI contributes to mitigating inequity. This study also examines whether and how individual characteristics are associated with such perceptions in these racial groups, as well as find an association between the perceptions and general attitude toward AI. As AI plays an increasingly important role in our society, these findings reveal racial disparity in perceived effectiveness and equity of AI, along with relevant factors. Overall, this study speaks to racial inequity in the context of technology development, contributes to our understanding of different racial groups’ preferences and concerns about AI, and calls for a development of AI that benefits different groups more equally.