Are you willing to forgive AI? Service recovery from medical AI service failure
| DOI | https://doi.org/10.1108/IMDS-12-2021-0801 |
| Published date | 16 August 2022 |
| Date | 16 August 2022 |
| Pages | 2540-2557 |
| Subject Matter | Information & knowledge management,Information systems,Data management systems,Knowledge management,Knowledge sharing,Management science & operations,Supply chain management,Supply chain information systems,Logistics,Quality management/systems |
| Author | Aihui Chen,Yueming Pan,Longyu Li,Yunshuang Yu |
Are you willing to forgive AI?
Service recovery from medical AI
service failure
Aihui Chen and Yueming Pan
College of Management and Economics, Tianjin University, Tianjin, China
Longyu Li
School of Economics and Management, Tongji University, Shanghai, China, and
Yunshuang Yu
College of Management and Economics, Tianjin University, Tianjin, China
Abstract
Purpose –As an emerging technology, medical artificial intelligence (AI) plays an important role in the
healthcaresystem. However,the service failureof medical AI causessevere violationsto user trust. Differentfrom
other servicesthat do not involve vital health, customers’trust toward the serviceof medical AI are difficultto
repair after service failure. This study explores the links among different types of attributions (external and
internal),service recovery strategies (firm,customer, and co-creation), and servicerecovery outcomes (trust).
Design/methodology/approach –Empirical analysis was carried out using data (N5338) collected from a
233 scenario-based experiment. The scenario-based experiment has three stages: service delivery, service
failure, and service recovery. The attribution of service failure was divided into two parts (customer vs. firm),
while the recovery of service failure was divided into three parts (customer vs. firm vs. co-creation), making the
design full factorial.
Findings –The results show that (1) internal attribution of the service failure can easily repair both affective-
based trust (AFTR) and cognitive-based trust (CGTR), (2) co-creation recovery has a greater positive effect on
AFTR while firm recovery is more effective on cognitive-based trust, (3) a series of interesting conclusions are
found in the interaction between customers’attribution and service recovery strategy.
Originality/value –The authors’findings are of great significance to the strategy of service recovery after
service failure in the medical AI system. According to the attribution type of service failure, medical
organizations can choose a strategy to more accurately improve service recovery effect.
Keywords Artificial intelligence, Service recovery, Trust repair, Medical AI, Co-creation recovery
Paper type Research paper
1. Introduction
Trust is crucialto human behavior and interaction.As the sociologist NiklasLuhmann said, it
constitutes“a mechanismto reduce social complexity”.Trust is needed in financial,medical and
other servicefields (Goles et al.,2009).In the field of medicine, the developmentof AI promotes
its significantadvantage over human doctors (Shi et al.,2021). However, trust in medical AI is
verydifficult to build. The servicefailure is inevitableeven though the probabilityof occurrence
is extremely low. Butcases on the failure of medical AI may have seriousnegative effects on
patient experien ce (Quinn et al.,2021). A wrong diagnosis makes the patients exposed to life-
threatening illness and brings them greatpsychological pressure (Kelly et al.,2019).
The most significant problem on the service failure of medical AI still exists before the
promotion of the AI system. The trust toward the medical AI system is difficult to repair after
the service failure. Customers are always unwilling to rebuild the trust even though the AI
system is more efficient than human doctors in disease diagnosis (Miller and Brown, 2018).
IMDS
122,11
2540
Funding: This work was supported by the Tianjin Philosophy and Social Science Planning Project
(TJGL21-003).
The current issue and full text archive of this journal is available on Emerald Insight at:
https://www.emerald.com/insight/0263-5577.htm
Received 31 December 2021
Revised 19 April 2022
19 July 2022
Accepted 27 July 2022
Industrial Management & Data
Systems
Vol. 122 No. 11, 2022
pp. 2540-2557
© Emerald Publishing Limited
0263-5577
DOI 10.1108/IMDS-12-2021-0801
According to counterfactual thinking, customers consider how the medical AI system could
have behaved (Hengstler et al., 2016). The way medical AI deals with the problem is still a
“black box”to customers, which makes it difficult for customers to understand how it will
diagnose. It is the reason for that trust in medical AI is more difficult to obtain than in other
service fields. People would prefer to go to the hospital rather than use the medical AI system,
even if its service has been recovered and the latter is much more convenient. As a result of
the problems above, repairing trust is significant to the application of the medical AI system.
At the same time, we face a significant gap in the knowledge of service recovery of medical
AI. Existing studies on service recovery mainly focus on enterprise organization services and
product services (Suparna Biswas et al., 2020;Bagherzadeh et al., 2020,Lv et al., 2022). A few
studies on service recovery in the field of medical AI are present. The attribution theory is the
most relevant theory used on the effects of service recovery in recent studies. However, no
study has examined the service recovery outcomes in medical AI using the attribution theory.
Based on the involvement of firms and customers during the service recovery process, service
recovery strategies can be divided into three categories: firm recovery, customer recovery,
and co-creation recovery (Dong et al., 2008). However, how the three service recovery
strategies and their interaction with attributions impact the service recovery outcomes in the
context of medical AI remains unknown.
Therefore, our research questions are as follows:
(1) (How do different types of attributions (internal and external) of service failure affect
the outcomes of service recovery of medical AI in trust?
(2) (How do different types of service recovery strategies (firm recovery, customer
recovery, and co-creation recovery) affect the outcomes of service recovery of medical
AI in trust?
(3) (What is the interaction effect between different attributions and service recovery
strategies?
To address these questions, we developed a research model to elaborate the relationships
among attribution types, service recovery strategies and repaired trust. To test the
hypotheses, we conducted a scenario-based experiment (N5338) incorporating three stages:
initial service, service failure, and service recovery. Attribution theory was used to divide the
attribution into customer (internal) and firm (external), while service recovery theory was
used to divide the recovery method into customer, firm, and co-creation. We found that co-
creation recovery has a greater positive effect on AFTR, while firm recovery is more effective
on CGTR. Furthermore, a series of interesting conclusions were found in the interaction
between the customers’attribution and service recovery strategy. We also concluded that the
effect of repair is better in internal attribution than in external attribution. Co-creation is the
best among the three different recovery methods.
The findings of this research provide significant contribution for theory and practice.
First, we focus on the field of medical AI service failure and recovery and verify the
applicability of the attribution theory in medical AI. Second, by collecting data from the same
participants in three-stage scenario-base experiments, this study shows how trust (cognitive
and affective) change in different attributions and service recovery strategies. Third, by
analyzing the interaction between attributions and service recovery strategies, we contribute
a variety of specific service recovery strategies to operators for different outcomes.
2. Theoretical background and hypotheses development
Medical AI stores medical records in digital format and uses intelligent technology for patient
examination. It can provide personalized medical plans and treatments based on the specific
Service
recovery from
medical AI
failure
2541
To continue reading
Request your trial