Algorithm of Trust: The Causal Impact of AI Personalization on Consumer Adoption in Mobile Messaging Payments
DOI:
https://doi.org/10.71465/fbf718Keywords:
Artificial Intelligence, Personalization, Consumer Trust, Mobile Payments, Difference-in-DifferencesAbstract
The convergence of artificial intelligence and mobile messaging platforms has created a new paradigm for digital financial services, yet the mechanisms through which AI-driven personalization influences consumer adoption of payment features remain poorly understood. This paper investigates the causal impact of personalized AI recommendations on user adoption of mobile messaging payment services, with a specific focus on the mediating role of consumer trust. We leverage a natural experiment from a major messaging platform's phased rollout of an AI payment assistant to implement a difference-in-differences research design. Using a panel dataset of fifteen thousand users observed over eighteen months, we estimate the treatment effect of AI personalization on payment adoption rates while accounting for heterogeneous responses across user segments. Our findings reveal that AI personalization significantly increases payment adoption by twenty-three point four percent, with trust-based mechanisms explaining approximately forty-one percent of this total effect. Critically, we uncover substantial heterogeneity in treatment effects: users with prior trust in the platform exhibit adoption increases of thirty-one point two percent, while those with low initial trust show muted responses of only eight point seven percent, suggesting that AI personalization amplifies existing trust rather than building it de novo. We also find that recommendation transparency moderates the effect, with explainable AI features enhancing adoption among skeptical users. These results provide causal evidence for the algorithm of trust hypothesis, demonstrating that AI personalization functions through trust channels and that its effectiveness is conditional on pre-existing user-platform relationships. The findings offer actionable insights for designing AI systems that foster financial inclusion while managing the heterogeneous trust dynamics inherent in digital payment adoption.
Downloads
Downloads
Published
Issue
Section
License
Copyright (c) 2026 Yi Han (Author)

This work is licensed under a Creative Commons Attribution 4.0 International License.