Figure 1
The conceptual flow diagram consists of two side-by-side panels labeled “(a) Human-based Algorithm Aversion” on the left and “(b) AI-based Algorithm Aversion” on the right. In the left panel, a dashed boundary at the top separates advisor types. On the far left, a rectangle labeled “A I Advisor” contains a box labeled “A I Advice”. To its right, a dashed boundary labeled “Human Advisor” contains two boxes arranged horizontally, labeled “Self Advice” and “Human Advice”. Solid downward arrows from “A I Advice”, “Self Advice”, and “Human Advice” converge into a central rectangle labeled “Human Advisee’s Decision”. From this decision box, three solid downward arrows extend to a dashed boundary at the bottom labeled “Dependent Variable Measures”, containing three horizontally arranged boxes labeled “Advisor Choice: Use A I Advisor or Other (Human) Advisor”, “Weight of Advice: How much to use Human or A I Advise”, and “Ratings: Human or A I Advisor”. In the right panel, a dashed boundary at the top again separates advisor types. On the left side of this boundary, a rectangle labeled “A I Advisor” contains two horizontally arranged boxes labeled “(External) A I Advice” and “Self Advice”. To the right, a dashed boundary labeled “Human Advisor” contains a single box labeled “Human Advice”. Solid downward arrows from “(External) A I Advice”, “Self Advice”, and “Human Advice” converge into a central rectangle labeled “A I Advisee’s Decision”. From this decision box, three solid downward arrows lead into a dashed boundary at the bottom labeled “Dependent Variable Measures”, containing three boxes labeled “Advisor Choice: Use A I Advisor or Other (Human) Advisor”, “Weight of Advice: How much to use Human or A I Advise”, and “Ratings: Human or A I Advisor”.Conceptual framework contrasting human and AI-based algorithm aversion. Source: Authors’ own work