Risk and trust perceptions of the public of Artificial Intelligence applications

This paper describes a study on the perceived risk and trust of members of the general public regarding artificial intelligence applications. It assesses whether there is a difference in the perceptions of risk and trust in artificial intelligence expressed by the general public compared with those studying computer science in higher education. We define the general public as people having no specific level or specialist knowledge of AI yet with a high stake as potential users of AI systems on a regular basis with or without their knowledge. In the study, participants engaged in an AI debate on topical news articles at a public national science museum event and a University in the UK and completed a questionnaire with two sections: their assessment of trust and risk of an AI application based on a topical news story, and a set of general opinion questions on AI. Results indicate that in specific applications there is a significant difference of opinion between the two groups with regards to risk. Both groups strongly agreed that education in how AI works was significant in building trust