Factors Driving Individuals’ Usage Intention of Artificial Intelligence (AI) Assistants in E-commerce: Perspectives of Users and Non-Users
Date
2024
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
University of Technology Sydney
Abstract
The ongoing revolution of e-commerce has brought about significant transformations in the
global retail landscape, redefining how consumers interact with online platforms. In response
to this transformative trend, businesses increasingly adopt and integrate artificial intelligence
(AI) technologies, particularly AI assistants. AI assistants have gained significant traction to
enhance customer engagement, improve personalised experience, and streamline various
aspects of the e-commerce process. Companies across diverse industries and geographical
regions have recognised the potential of AI assistants in fostering more profound connections
with customers, providing real-time support, and bolstering sales through intelligent
recommendations. Consequently, investment in AI research and development has surged,
leading to remarkable advancements in AI assistants’ features and functionalities. Despite the
growing interest of the scientific community and business stakeholders in the topic, scholarly
research on the factors influencing e-commerce consumers’ attitudes and intentions toward
using AI assistants is still limited and provides contradictory evidence regarding some factors.
Moreover, no comparative studies in the e-commerce context empirically investigated the
attitudes of non-users and users toward AI assistant use. Also, several consumers'
demographics have been excluded from prior research, with no previous empirical research on
AI assistant use across different cultural backgrounds.
For these reasons, the study aimed to comprehend the factors influencing consumers'
behavioural intention to utilise AI assistants and to recognise the significant user differences
based on multiple perspectives. This study employed a unique research model based on the
technology acceptance model. It extended it with external factors of AI assistants’ capabilities
that still need to be tested together in AI assistant adoption for e-commerce consumers. This
research conducted a mixed-method approach. In the first phase (Phase A), a quantitative
method was employed to investigate the relationships between the constructs in the study
model, and the Partial Least Square Structural Equation Modelling (PLS-SEM) and several
statistical techniques were adopted. Furthermore, to account for cross-cultural differences and
identify potential variations in usage intentions towards using AI assistants between Eastern
and Western cultures, a multi-group analysis (MGA) was conducted. In the second phase
(Phase B), a qualitative approach was conducted by applying machine learning and natural
language processing techniques to analyse reviews of the Louis Vuitton brand's e-commerce
applications. The objective of this stage was to obtain supporting evidence for the results of the
VI
first study and to gain deeper insights into consumer attitudes and experiences. Subsequently,
the results were integrated to provide multiple insights to answer the research questions and
strengthen the findings.
This study has confirmed some previous studies' results and provided new findings. The
attitude factor was the significant predictor of the intention to use AI assistants in non-users
and users, with a direct and positive effect. Perceived usefulness was found to be the
statistically significant predictor of attitudes in both non-users and users of AI assistants. The
additions to the original TAM model, specifically incorporating interactive communication and
personalisation, were statistically significant predictors of the attitudes of non-users and users
to use AI assistants with positive effects. In contrast, perceived ease of use was a nonsignificant predictor of the non-users’ attitudes and positively impacted the users’ attitudes
towards using AI assistants. Furthermore, no significant differences existed in the relationships
among the primary factors influencing the intention to utilise AI assistants in e-commerce when
comparing Western and Eastern cultural groups.
This study contributes to both theory and practice by extending the TAM model with two
external factors enabling the assessment of the factors affecting the intention to use AI
assistants from consumer, social, and marketing perspectives and providing new empirical data
on this topic in technology adoption studies. The study also enables further research on this
topic and comparing study results, thus improving understanding of the phenomenon. It also
provides various e-commerce practitioners with valuable information and recommendations
regarding AI assistant use, enabling them to make better decisions in developing and
implementing AI assistant technologies.
Description
Keywords
Artificial Intelligence, AI assistants, chatbots, E-commerce, E-commerce applications, Technology Acceptance Model, AI Chatbots, PhD research, Adoption