Artificial Intelligence (AI) Literacy: Development of a Measure of Knowledge and Trust

Background: Differences in general knowledge and understanding of the benefits and dangers regarding artificial intelligence (AI) are becoming important topics in technology fields. Key differences in understanding of AI could influence how people from diverse backgrounds trust and interact with AI.  Factors that may affect an individual’s perception of AI may include their cultural background, gender, race, and language.  This study evaluated such potential differences with a particular focus on those who may have very different levels of knowledge and/or trust in AI based on their race, culture, and gender.

Method: To answer these research questions, this study developed an online self-report measure designed to evaluate an individual’s regular use and familiarity with, and knowledge of, AI systems, and their level of trust in these systems.  The AI literacy measure consisted of 28 questions, some of which were multi-component, which were categorized into a few conceptual domains: the participant’s knowledge about AI, their regular AI use, their understanding of the AI involved, and their trust of such AI systems. It was hypothesized that there would be an inverse linear relationship between the two key variables; as a person’s general knowledge about AI increases, their trust will decrease.

Results and Discussion: One hundred and seventeen participants completed the AI Literacy measure on Qualtrics, most of the participants identified as African American and female. SPSS was used to evaluate their results.  Results from various components will be presented highlighting similarities and differences in responses between races and genders.  In addition, correlation analyses addressed the primary hypothesis.  These results are considered to be a starting point for developing better measures designed to better understand the level of confidence in which humans have in AI. Additionally, these results will contribute to the present research on AI ethics and the development of more equitable technologies free of bias.