ObjectiveWe explored the Investment construct to quantify human utilization of AI and examined respect, trust, workload, and self-confidence as factors influencing utilization within the Investment Framework.BackgroundThe Investment construct is a new approach proposed to quantify human utilization of AI-enabled agents based on the measurable opportunity costs incurred by humans. An experimental study was conducted to operationalize the construct and examine its relationships with influencing factors.MethodThirty participants performed a maintenance task while an AI assisted them. Participants were split into three groups with different allocated completion times to simulate different task-load levels, under the assumption that a lower task load would lead to higher Investment. We collected typing data as the objective measure of Investment, and responses to a six-item custom scale as the subjective measure. Additionally, we collected data for the influencing factors listed above.ResultsThere was a significant difference between groups in terms of both Investment measures, and Investment was higher with a lower task load, as expected. Furthermore, these two Investment measures correlated. We also revealed a correlation between respect and the subjective measure of Investment. No other significant correlations were revealed.ConclusionThis study presents the first empirical evidence of AI utilization conceptualized through incurred opportunity costs within the Investment Framework, explored using two independent measures and influencing factors.ApplicationThe Investment Framework can guide the design of effective human AI interaction by helping to balance human effort and AI benefit, prevent underutilization, and enhance both team performance and AI improvement.
Ulusoy et al. (Tue,) studied this question.