Paper Type

Short

Paper Number

PACIS2026-1319

Description

As agentic AI is increasingly introduced into organisational workflows, delegating execution authority to AI becomes central in hybrid human–AI collaboration where autonomy is shared between humans and agentic systems. While Information Systems research recognises the power dynamics involved in sharing autonomy, limited attention has been given to how AI autonomy configurations shape users’ psychological responses during interaction episodes and influence their willingness to delegate decisions in future interactions. This study develops a situational power calibration model explaining how AI autonomy configurations influence users’ perceived decisional power and subsequently inform delegation intention. Drawing on the approach–inhibition theory of power, we propose dual psychological pathways through which autonomy configurations activate reward–trust or control threat–reactance responses. By introducing interaction-level power calibration, the study extends existing delegation frameworks and highlights how autonomy configurations can influence trust, resistance, and delegation behaviour in human–agentic AI collaboration.

Comments

03-EthicsSocietalImpact

Share

COinS
 
Jul 5th, 12:00 AM

Power Calibration and Delegation in Human–Agentic AI Interaction: Psychological Pathways

As agentic AI is increasingly introduced into organisational workflows, delegating execution authority to AI becomes central in hybrid human–AI collaboration where autonomy is shared between humans and agentic systems. While Information Systems research recognises the power dynamics involved in sharing autonomy, limited attention has been given to how AI autonomy configurations shape users’ psychological responses during interaction episodes and influence their willingness to delegate decisions in future interactions. This study develops a situational power calibration model explaining how AI autonomy configurations influence users’ perceived decisional power and subsequently inform delegation intention. Drawing on the approach–inhibition theory of power, we propose dual psychological pathways through which autonomy configurations activate reward–trust or control threat–reactance responses. By introducing interaction-level power calibration, the study extends existing delegation frameworks and highlights how autonomy configurations can influence trust, resistance, and delegation behaviour in human–agentic AI collaboration.