LAUSR

Introduction The use of artificial intelligence (AI) for national defense is a matter of high societal significance and ongoing public discourse, but very little is known about public acceptance of AI in defense contexts. Currently, there is no reliable and valid measure of attitudes towards AI in defense, and more general attitudinal measures on AI use are unlikely to capture relevant perceptions and opinions. A measure was therefore developed for the assessment of Attitudes towards AI in Defense (AAID), and this work presents the initial validation of this scale. Methods A total of 1,590 participants (aged 19-75, M = 45.7, SD = 16.1) completed a self-report questionnaire which included an initial item pool of 29 attitudinal statements related to the use of AI in defense. An additional general attitude towards AI scale was also included to assess the concurrently validity of the AAID scale. The AAID underwent initial statistical validation via exploratory factor analysis and confirmatory factor analysis to test the underlying structure of the newly developed scale. Results Items reduction and exploratory factor analysis resulted in a final scale consisting of 15 items. A final two factor solution explained 42.52% of the variance (Factor 1 = 22.35%, Factor 2 = 20.17%). Factor 1 was termed “Positive outcomes” and reflected the potential and anticipated consequences of implementing AI in defense. Factor 2 was termed “Negative Outcomes” and reflected the potential negative outcomes for AI in defense. The scale also exhibited acceptable internal reliability and current validity. Discussion The newly developed AAID presents a new measurement tool which has the capacity to assess current attitudes towards AI in defense. Such work is essential if further developments in AI in defense are to continue with the support of the public. However, the work also notes that there are some key concerns and barriers that could block further developments in the area, with further work needed to explore how such anxieties are driven by narratives related to the topic.