Jolene Cox

Suicide is a leading cause of premature mortality worldwide, making suicide prevention a global public health priority. As more Artificial Intelligence (AI)-based suicide prevention interventions are being developed and implemented, it is important to study acceptance of these AI systems. The present study aimed to investigate the factors that predict acceptance of current and future AI systems for suicide prevention and the perceived risks and benefits of these AI systems. Individuals from the Australian public were invited to participate in an online survey, which included six hypothetical scenarios of current AI systems (Artificial Narrow Intelligence [ANI]) and future advanced AI systems (Artificial General Intelligence [AGI]) for suicide prevention. Participants evaluated these scenarios on five technology acceptance factors (performance expectancy, effort expectancy, social influence, facilitating conditions, trust) and elaborated on its perceived risks and benefits. Performance expectancy, social influence, and trust predicted acceptance of ANI systems, but only trust predicted acceptance of AGI systems. Overall, the level of acceptance was higher for ANI systems than for AGI systems. Several perceived risks (e.g., risks to mental healthcare, distrust in AI, threats to humanity) and perceived benefits (e.g., benefits to mental healthcare, trust in AI, human help-seeking) were identified. AI systems represent potential avenues for effective suicide prevention. However, to ensure acceptance, AI systems for suicide prevention must be developed in a way that is safe, reliable, and trustworthy.

Previous research has demonstrated that anxious individuals attend to negative emotional information at the expense of other information. This is commonly referred to as attentional bias. The field has historically conceived of this process as relatively static; however, research by [Zvielli, A., Bernstein, A., & Koster, E. H. W.. Dynamics of attentional bias to threat in anxious adults: Bias towards and/or away? PLoS ONE, 9, e104025; Zvielli, A., Bernstein, A., & Koster, E. H. W.. Temporal dynamics of attentional bias. Clinical Psychological Science, 3, 772–788.], and others, challenges this assumption by demonstrating considerable temporal variability in attentional bias amongst anxious individuals. Still, the mechanisms driving these temporal dynamics are less well known. Using a modified dot-probe task, the present study examined the impact of two relevant contextual variables- affective valence and trial repetition. Affective context was instantiated by the presentation of...