“The more you know, the less you fear.”: A quantitative study on the effects of AI literacy, AI self-efficacy and perception of AI uncanniness on AI anxiety amongst adults.

With the rapid growth of Artificial Intelligence across various sectors, researchers have found an increase in fear and anxiety amongst the general population. Much of the research found today focuses on acceptance of technology models and analysis of effectiveness of these technologies. There is also a focus on whether anxiety surrounding technology leads to a decrease in acceptance. However, there is a critical lack of research into prevention of anxiety towards Artificial Intelligence. This thesis aims to tackle this research gap by analysing what variables can in fact lower AI anxiety amongst the general public. The dependent variables chosen for this thesis were AI literacy, AI self-efficacy, perception of AI uncanniness and innovativeness was chosen as a moderating factor. For this, an online survey was distributed that measured AI literacy, AI self-efficacy, AI uncanniness perceptions, innovativeness and AI anxiety. After collecting 154 valid results through distributing the survey on survey deploying sites such as Prolific and Surveyswap, the data was imputed into IBM SPSS software and was used to test various regressions and moderator relationships between the variables. In particular, it was found that participants with higher AI literacy are more likely to have lower AI anxiety. This relationship is also strengthened by the presence of high innovativeness amongst the participants. It is also found that participants with a negative perception of AI uncanniness are more likely to have higher AI anxiety. It is also important to note that lowering the negative perception towards AI uncanniness is important in lowering AI anxiety. These findings are useful in a range of various market sectors from education policy to marketing and AI developers.