Paper Number

ICIS2025-1570

Paper Type

Short

Abstract

This study examines paradoxical effects of user controllability in AI-driven financial services using 312 weeks field data from a South Korean robo-advisor. We investigate how decision control (approving algorithm's rebalancing) and process control (user-initiated rebalancing) influence user behavior and outcomes. Our findings reveal that decision control reduces attention (~34%) and deposits, yet increases risk-adjusted returns (+4.4pp). Adding process control raises attention (~29%) and restores deposit activity but erodes returns (−3.9pp). While reduced attention under decision control improves portfolio performance by minimizing counterproductive manual interventions, it significantly decreases platform engagement metrics. Conversely, increased attention under process control drives engagement but triggers frequent manual rebalancing that systematically undermines algorithmic performance. This illuminates a fundamental tension between algorithm performance and user engagement, revealing trade-offs for platforms balancing financial outcomes with user retention in human-AI collaboration.

Comments

22-FinTech

Download

Share

COinS
 
Dec 14th, 12:00 AM

Controllability in Algorithmic Decision-Making: Field Evidence from a Robo-Advisor

This study examines paradoxical effects of user controllability in AI-driven financial services using 312 weeks field data from a South Korean robo-advisor. We investigate how decision control (approving algorithm's rebalancing) and process control (user-initiated rebalancing) influence user behavior and outcomes. Our findings reveal that decision control reduces attention (~34%) and deposits, yet increases risk-adjusted returns (+4.4pp). Adding process control raises attention (~29%) and restores deposit activity but erodes returns (−3.9pp). While reduced attention under decision control improves portfolio performance by minimizing counterproductive manual interventions, it significantly decreases platform engagement metrics. Conversely, increased attention under process control drives engagement but triggers frequent manual rebalancing that systematically undermines algorithmic performance. This illuminates a fundamental tension between algorithm performance and user engagement, revealing trade-offs for platforms balancing financial outcomes with user retention in human-AI collaboration.

When commenting on articles, please be friendly, welcoming, respectful and abide by the AIS eLibrary Discussion Thread Code of Conduct posted here.