Using artificial intelligence (AI)? Risk and opportunity perception of AI predict people’s willingness to use AI

AbstractSurveys worldwide show that the public perceives artificial intelligence (AI) as a double-edged sword: A risk and an opportunity. However, how this ambiguous perception of AI is related to people's willingness to use AI-based applications has yet to be investigated. To this end, two online experiments were conducted, including two samples, N = 246 and N = 495 (quota-sample, representative for age and gender). As hypothesized, people's risk-opportunity perception of AI applications correlated positively with the probability of using AI. Exploratory analyses indicated that people's willingness to use AI significantly depended on the context of AI use (medicine vs. transport vs. media vs. psychology). This research expands existing behavioral research by investigating ambiguous and not solely risk-taking behavior for different AI application contexts. Study results motivate the investigation of causal-effect relations and underline the need to understand risk and opportunity perception stability across different contexts of AI use.Keywords: Risk perceptionopportunity perceptionartificial intelligencebehaviorconfidence Ethical approvalAPA ethical standards were followed in the conduct of both studies reported in this article and informed consent was collected from the participants at the beginning of the study. Both studies were approved by the ethics committee of the Leibniz-Institut für Wissensmedien, Tübingen.Authors' contributionsNS, IB, JB, and MH developed the research idea. NS administered the whole project and developed the study concepts. Both, RS and NS developed the methodology of the studies and analyzed the data. RS was responsible for data collection and data visualization. RS and NS wrote the original draft. All authors were responsible for reviewing and editing the original draft. All authors approved the final version of the manuscript for submission.Disclosure statementNo potential conflict of interest was reported by the author(s).Data availability statementData of both studies are freely accessible under http://dx.doi.org/10.6084/m9.figshare.20589597 (Schwesig and Said 2021, dataset). The analysis code (R) that produces all results and figures of this article are available at http://dx.doi.org/10.6084/m9.figshare.20589597. Before data collection, both experimental studies were preregistered: https://aspredicted.org/CVC_CRQ (study 1) and https://aspredicted.org/SR3_8Q3 (study 2).Notes1 Note, that we preregistered that people's AI knowledge works as a moderator on the association of risk and opportunity perception and behavior towards AI as hypothesis 2 for the second study.2 Note, that for study 2 there was a significant main effect of knowledge when we did not control for age, gender, and education, and entering knowledge as main effect only: Χ2Study2(2) = 4.36, p = .037, OR =7.42, 95% CI [1.12, 49.37].3 Note, that for study 2 there was a significant main effect of confidence when we did not control for age, gender, and education, and entering confidence as main effect only: Χ2Study2(2) = 5.12, p = .024, OR =1.22, 95% CI [1.02, 1.44].Additional informationFundingData collection was funded by internal funds of the Leibniz-Institut für Wissensmedien (Tübingen, Germany).