The Impact of Artificial Intelligence Technologies on Cross-Border

e-Commerce Supply Chain Resilience

Jinxiang Chen

School of Management, Hefei University of Technology, Xuancheng 242000

，

China

Keywords: Artificial Intelligence, Cross-Border e-Commerce, Supply Chain Resilience.

Abstract: Against the backdrop of heightened global economic uncertainty, cross-border e-commerce supply chains

face multiple risks, including geopolitical tensions, logistics disruptions, and demand fluctuations. The rapid

evolution of digital technologies, particularly artificial intelligence, has introduced both opportunities and

challenges for optimizing supply chain resilience in volatile environments. Meanwhile, the post-pandemic

global economic landscape has further accelerated the digital transformation of cross-border trade, making

AI-driven solutions increasingly critical for adaptive supply chain management. This paper focuses on the

impact mechanism of AI technology on the resilience of cross-border e-commerce supply chains, specifically

addressing the unique characteristics of their cross-border nodes and complex structures. Using a theoretical

research approach, an analytical framework is constructed from three dimensions: national policies, pandemic

impacts, and supply chain integration. By combining case studies and literature reviews, the paper

systematically explores the application logic and limitations of AI technology in scenarios such as policy

analysis, demand forecasting, and logistics optimisation. The research findings indicate that AI plays a

significant role in enhancing the resilience of cross-border e-commerce supply chains, but its effectiveness

remains constrained by technical applicability and environmental complexity.

1 INTRODUCTION

In the digital economy era, international trade has

become more important through cross-border e-

commerce due to the intensified global economic

uncertainty. Cross-border e-commerce supply chain

faces multiple risks such as geopolitical conflicts,

logistics disruption and demand fluctuations, and the

resilience construction has become a key proposition

to ensure the security and sustainable development of

the industrial chain. In recent years, research on the

impact of supply chain resilience has been analysed

mainly from the perspectives of national policy (Liu

& Qin, 2025), epidemic (Agca et al., 2023), and

supply chain integration (Qi et al., 2023). "Supply

chain resilience" was conceptualized in 2003 without

formal definition. The first formal definition,

established in 2004, describes it as a supply chain's

capability to recover from perturbations and optimise

its subsequent state (Gao et al., 2021). Other

definitions have been proposed, but the most widely

used characterizes resilience as a supply chain's

https://orcid.org/0009-0001-7031-2582

adaptive capacity to prepare for anticipated risks,

respond swiftly to disruptions, and efficiently recover

from them (Ponomarov & Holcomb, 2009).

At this stage, regarding the impact of AI

technology on supply chain resilience, some scholars

have studied the impact of AI technology on supply

chain resilience of automotive manufacturing

companies (Liu, 2025), while others have explored

the impact of AI on supply chain resilience based on

a specific timing and considering that AI enhances the

supply chain resilience by improving the visibility,

risk, sourcing, and distribution capabilities(Modgil et

al., 2022) These studies have laid a solid foundation

for revealing the correlation between AI technology

and supply chain resilience, but there are still

limitations and shortcomings in analysing the

relationship in depth. Firstly, the majority of the

literature that exists examines the impact of AI

technology on supply chain sustainability from an

overall perspective, but rarely explores the influence

of a specific factor on supply chain sustainability

caused by AI technology; secondly, compared with

300

Chen, J.

The Impact of Artiﬁcial Intelligence Technologies on Cross-Border e-Commerce Supply Chain Resilience.

DOI: 10.5220/0014351800004718

Paper published under CC license (CC BY-NC-ND 4.0)

In Proceedings of the 2nd International Conference on Engineering Management, Information Technology and Intelligence (EMITI 2025), pages 300-305

ISBN: 978-989-758-792-4

the supply chain of traditional manufacturing

enterprises, cross-border e-commerce supply chain

has many nodes, cross-border nodes, and the

characteristics of complex structure(Liu & Yang,

2024) , yet existing research has not sufficiently

addressed the similarities and differences between

cross-border e-commerce enterprises and traditional

manufacturing enterprises in their use of AI

technology. Adopting a theoretical research

framework, this paper analyses AI technology impact

on cross-border e-commerce supply chain resilience,

and puts forward the suggestions based on the

influencing factors.

2 IMPACT STUDIES IN

DIFFERENT CONTEXTS

2.1 Research in the Context of National

Policy

National policies often have a dual impact on supply

chain resilience. On the one hand, policies that

promote mutual benefit and cooperation tend to

enhance supply chain stability and resilience. Take

the Belt and Road Initiative as an example: its core

focus is not only on political trust and trade

facilitation but also on supporting the development of

critical logistics infrastructure along the route. China

has collaborated with relevant countries to build over

a hundred smart logistics hubs and overseas

warehouses, leveraging big data analysis and

automated equipment, have improved warehouse

operation efficiency and optimised logistics routes on

a global scale. The construction and operation of

land-based transport routes such as the China-Europe

Railway Express have significantly reduced cross-

border transport times, improved logistics efficiency,

and enhanced the ability to select alternative transport

routes in the event of local transport disruptions

caused by sudden disturbances, thereby objectively

reducing the risk of supply chain disruptions from a

physical perspective. Additionally, government-

driven initiatives such as cross-border e-commerce

comprehensive pilot zones and free trade port

development have substantially simplified customs

clearance procedures and offered tariff exemptions

for specific goods, significantly reducing enterprises'

compliance costs and time costs, thereby significantly

enhancing the agility of supply chains in responding

to fluctuations in international markets. On the other

hand, national policies with trade protectionist

characteristics often lead to adverse effects such as

supply chain disruptions. As China has made

breakthroughs in core technology fields such as

artificial intelligence, high-speed rail, and chips,

Western countries led by the United States are

accelerating the “de-Chinaisation” of supply chains,

attempting to strategically suppress China through

technological ’decoupling’. The U.S.‘s technology

“decoupling” policy towards China has forced cross-

border e-commerce companies to reconfigure their

supplier networks, increasing the risk of supply chain

disruptions in the short term. However, Liu (Liu &

Qin, 2025) and others have also pointed out that while

the U.S.’s technology ‘decoupling’ policy may

initially impact businesses, it has also spurred strong

momentum for independent R&D. Additionally,

cross-border e-commerce companies are being

compelled to adopt technologies such as the Internet

of Things, artificial intelligence, and big data analysis

to enhance supply chain efficiency. These companies

are also optimising their partner networks, enhance

supply chain concentration, reduce supply chain

cooperation complexity, and adopt strategies such as

concentrating resources, strengthening core

capabilities, and rapidly adjusting supply chain

layouts to reduce inventory costs and optimise supply

chain processes (Christopher & Peck, 2004). This will

help solidify strategic partnerships with key suppliers,

logistics service providers, and business partners,

enhance supply chain resilience and corporate risk

resistance capabilities, and apply technologies such as

blockchain and cloud computing to streamline

intermediate processes and eliminate redundant and

inefficient complexities.

It is important to address the impact of different

types of national policies on cross-border e-

commerce enterprises and to move them in a direction

that is conducive to improving supply chain

resilience. Through the research of Venkatesh

Shankar (Shankar & Parsana, 2022) and others, it is

found that Natural Language Processing (NLP)

technology is a very suitable technology to analyse

policy texts, determine the policy direction, and warn

the risk of supply chain disruption in advance.

Through text analysis, researcher can perform

keyword extraction, sentiment analysis, topic

modelling and other operations on policy documents.

For example, researcher can use keyword extraction

to understand the topics discussed in the documents;

sentiment analysis to understand the emotional

tendencies expressed in the documents; and theme

modelling to identify the main ideas in the

documents. Sentiment analysis, which assesses the

emotional tendencies expressed in a text, is essential

for understanding the intent and impact of a policy

The Impact of Artiﬁcial Intelligence Technologies on Cross-Border e-Commerce Supply Chain Resilience

301

document. Libraries such as TextBlob and VADER

in Python can help a researcher perform sentiment

analysis. For example, a researcher can use TextBlob

to calculate the emotional tendencies expressed in a

document to understand the emotional colour of the

document. Entity recognition is an important task in

NLP that can help researcher to identify and

understand specific entities and concepts in policy

documents. SpaCy library in Python has a powerful

entity recognition feature that can be used to extract

entities such as names of people, places,

organisations, and so on from documents. However,

NLP techniques still face some challenges in policy

document processing, such as language specificity

and differences in cultural backgrounds, complex

sentences and semantic ambiguity. Future research

needs to further explore how to improve the accuracy

and generalisation of NLP techniques in policy

document processing.

In general, policy orientation uncertainty

increases enterprise risk exposure, and cross-border

e-commerce faces different national policies,

regulations, customs clearance rules and other

environments due to the fact that the two parties'

transaction subjects are located in different

international market environments; therefore, in the

face of the pressure of sudden changes in the

international economic and trade environment, cross-

border e-commerce needs to improve the efficiency

and accuracy of judging the policy orientation, yet

cross-border e-commerce, often due to the differences

in geographic cultures, is often not able to However,

cross-border e-commerce companies often cannot

accurately understand the precise intention of the

policy due to the differences in regional cultures, and

thus need the advantage of objectivity and cross-

culturalism of AI technology in text analysis and

processing to enable the enterprises to make more

accurate judgments, and the policy and technology

can form a benign circle here, where the policy

creates the conditions for the use of the technology

and the efficient application of the technology feeds

back to the achievement of the policy objectives,

which together contribute to a more resilient global

supply chain network that is more able to withstand

the winds and waves.

2.2 Research in the Context of the

Epidemic

Since the first case was declared at the end of 2019,

as of 11 May 2025, 777,825,189 COVID-19 cases

have been reported globally, and the number of deaths

has come to 7,095,903 (data from WHO). Supply

chains in almost all industries worldwide have been

severely disrupted by the COVID-19 outbreak, with

drastic shifts in demand triggered by the outbreak The

dramatic change in demand triggered by the epidemic

and the ad hoc policies introduced by governments in

response to the outbreak were the core triggers of the

supply chain disruptions.

Epidemic-induced surge in epidemic prevention

materials and stagnant sales of non-essential goods

are more significant in cross-border e-commerce, and

AI technology can improve supply chain agility by

using intelligent forecasting and demand

management. Shen (Shen & Sun, 2023) et al.'s

analysis of JD's case shows that JD's intelligent

forecasting platform achieves a dynamic response to

unexpected dynamic response to demand.

Meanwhile, in cross-border scenarios, such

technology can be further upgraded: first, AI can

integrate data such as overseas social media public

opinion, customs policy changes, and local logistics

timeliness to build demand prediction models for

different countries. For example, when an epidemic

breaks out in Europe and the US, the AI system warns

of a surge in demand in advance by crawling the

search volume of masks on Amazon, eBay and other

platforms in real time, and guides domestic stock

preparation. Secondly, for slow-selling products due

to the epidemic, JD launched promotional campaign

to reduce unhealthy inventory, while AI can analyse

the differences in demand in different markets, thus

effectively reducing backlog inventory, ensuring the

return of funds, and maintaining the stability of the

supply chain. Finally, AI can automatically adjust the

replenishment cycle of cross-border goods based on

the output of the prediction model, compressing the

traditional 6-8 week cross-border replenishment

cycle into 72 hours, effectively alleviating the

"bullwhip effect".

Epidemic-induced international capacity

shortages and border controls often lead to cross-

border e-commerce supply chain disruptions, and this

core pain point has been previewed in the JD Wuhan

regional distribution centre disruption. However, JD

used the network optimisation capability of the JD-

NetSIM system to respond quickly and solve the

supply chain disruption problem. For example, in

2020, when the US-China air transport shutdown

prevented some goods from being delivered, JD's AI

system automatically initiated alternatives, such as

shipping goods from Shenzhen port to Los Angeles

by sea, and then distributing them to Amazon's FBA

warehouse through the trucking paths calculated by

AI, which resulted in a delay of 48 hours in the

timeframe, despite the increase in cost. During the

EMITI 2025 - International Conference on Engineering Management, Information Technology and Intelligence

302

COVID-19 outbreak, JD launched its first contactless

"mini distribution station" on February 20, 2020.

These hubs met public health mandates by operating

with minimal on-site parcels and full personal

protective equipment for workers. Staff underwent

specialized contact-prevention training, daily health

reporting, and regular testing, with all operations

featuring remote monitoring. Finally, due to the need

to reduce human contact during the epidemic, the last

kilometre of transport is often not possible in areas

where the epidemic is more severe, seriously

affecting the delivery of goods to customers, at this

time JD's unmanned delivery technology can solve

the problem of supply chain disruption due to the last

kilometre problem, improve customer satisfaction

while maintaining the stability of the supply chain, JD

also works with the community and parcels in shops,

supermarkets and other venues safe deposit boxes,

property management offices, and security offices to

achieve non-contact delivery and achieve good

results.

Overall, after the COVID-19 outbreak, the global

supply chain faced the double impact of drastic

changes in demand and disruptions in international

logistics. Cross-border e-commerce companies use

AI technologies to significantly improve supply chain

resilience through two dimensions: first, intelligent

forecasting and agile response on the demand side;

and second, intelligent scheduling and contactless

fulfilment on the logistics side. These technologies

maintain supply chain stability and customer

satisfaction under fluctuating demand and capacity

shortage.

2.3 Research in the Context of Supply

Chain Integration

The focus of the studies mentioned in the previous

two paragraphs is on examining the adaptive capacity

of supply chain resilience to respond promptly and

recover from interruptions or disruptions, whereas

policy impacts belong to human factor disruptions

and epidemic impacts belong to natural factor

disruptions, therefore, this paragraph will examine

the supply chain's pre-preparedness for potential

emergencies to improve the resilience of the supply

chain. Supply chain integration centres on

synchronizing external partners like suppliers and e-

commerce platforms to fulfil customer requirements.

This contrasts with vertical integration, which

achieves coordination through direct ownership of

entities along the value chain. E-commerce platforms,

which are integrated online retailers that take the lead

in their supply chains, provide the ICT infrastructure

that facilitates co-operation with suppliers, and can

therefore co-ordinate the e-commerce platform's

supply chain without the need for the layered

governance that is common in traditional supply

chains. Since there is no distribution stage, e-

commerce platforms typically purchase all products

directly from suppliers to simplify their supply chain

structure and long-term contracts, compared to

traditional supply chains with loose structures.

Consequently, sourcing and logistics constitute the

backbone of e-commerce supply chains, where

synchronized data exchange and collaborative

forecasting serve as vital connectors between

platforms and their supplier networks. Qi (Qi et al.,

2023) et al. showed that dynamic information

coordination, collaborative planning and forecasting,

integrated logistics coordination and intelligent

procurement execution in supply chain integration

significantly reduce the recovery time after supply

chain disruptions due to the enhanced efficiency of

information flow and logistics synergy.

Artificial intelligence technologies inherently

have the potential to increase the speed of information

processing, optimise the quality of decision-making,

and enhance execution efficiency, making them ideal

tools for enhancing the effectiveness of supply chain

integration, and thus resilience. In terms of

information sharing, AI greatly exceeds the scope and

capabilities of traditional EDI. The NLP technology

mentioned above can scan global multilingual news,

social media and government announcements in real

time to provide early warning of risky events;

machine learning algorithms can integrate all types of

data required by the supply chain to build a more

accurate and comprehensive end-to-end supply chain

view, enabling cross-border e-commerce enterprises

to sense risks and adjust plans earlier and more

proactively. On Joint Programmes, AI-driven digital

twin technology can build virtual supply chain

models to simulate the dynamic response of the entire

supply chain network under different disruption

scenarios, significantly reducing delays and resource

wastage due to supply and demand mismatch in

practice. For logistics collaboration, AI offers

unprecedented dynamic optimisation capabilities.

Automated devices such as self-driving trucks,

drones, and autonomous mobile robots can flexibly

deploy and reorganise logistics networks when

disruptions occur; more critically, AI path

optimisation algorithms based on, for example, ant

colony algorithms and evolutionary algorithms can

continually adjust the optimal transport and

distribution routes, maximise the use of collaborative

The Impact of Artiﬁcial Intelligence Technologies on Cross-Border e-Commerce Supply Chain Resilience

303

logistics resources, and significantly compress

disruption response times.

However, in the field of procurement automation,

the application of AI faces severe challenges and may

even become a bottleneck for resilience. Qi 。、(Qi

et al., 2023) et al.'s study clearly indicates that

procurement automation significantly extends the

recovery time during disruptions. The core reason lies

in the fact that procurement AI systems reliant on

historical data and static rules are highly prone to

‘concept drift’ when faced with ‘black swan’ events,

meaning that the historical demand patterns learned

by the model become severely disconnected from the

drastically changed reality. For example, during the

early stages of the pandemic, replenishment models

based on past stable sales continued to generate a

large number of procurement orders, while in reality,

consumption stagnated and logistics were disrupted,

leading to severe inventory buildup of unsold goods,

shortages of essential items, and cash flow pressures.

This failure highlights the vulnerability of

algorithmic decision-making under extreme

conditions. Especially for cross-border e-commerce,

which is characterised by high demand volatility,

strong market sensitivity, and rapid product iteration,

the inherent inertia risks of procurement automation

systems may be further amplified.

Therefore, the ultimate impact of AI technology

on cross-border e-commerce supply chain resilience

is not simply facilitated or hindered, but highly

dependent on the specific way and maturity of its

application in integration practices.

3 CONCLUSIONS

This paper systematically analyses the impact

mechanism of AI technology on the resilience of

cross-border e-commerce supply chains from three

dimensions: national policies, the impact of the

pandemic, and supply chain integration. The study

finds that, in terms of policy adaptability, AI

technology can effectively analyse the complexity

and contradictions of policies across multiple

countries, assist cross-border e-commerce enterprises

in anticipating the risks of trade protectionism, and

address policy uncertainties by optimising supplier

networks and logistics routes, thereby shifting from

‘passive response’ to ‘proactive resilience building.’

In terms of pandemic emergency response, AI

achieves dynamic forecasting through multi-source

data fusion on the demand side, alleviating the

‘bullwhip effect’; on the logistics side, it breaks

through international transport capacity bottlenecks

through intelligent scheduling and unmanned

technology, ensuring supply chain continuity. In

terms of supply chain integration, AI significantly

enhances information sharing, joint planning, and

logistics coordination. However, procurement

automation may become a resilience weakness due to

the ‘concept drift’ issue, as its algorithms overly

reliant on historical data exhibit fragility during

extreme events. Overall, AI serves as the core driving

force for resilience building by enhancing the supply

chain's perceptiveness, responsiveness, and

resilience. However, its effectiveness is constrained

by the applicability of the technology to specific

scenarios and the complexity of the external

environment.

This study makes dual contributions to theory and

practice. In terms of theoretical significance, it breaks

through the limitations of the existing literature

focusing on traditional manufacturing industries,

reveals the specificities of cross-border e-commerce

supply chains and the differentiated needs for the

application of AI technology, and supplements the

theory of supply chain resilience with new evidences

from cross-border contexts. The proposed dialectical

framework of the relationship between AI and

resilience: technology is both an "enabler" and a "risk

point", deepening the understanding of the double-

edged sword effect of technology. In terms of

practical significance, it provides enterprises with

actionable AI tool paths, such as NLP models, a

framework for fusion of multi-source data for demand

forecasting, and principles for designing intelligent

scheduling solutions for logistics disruptions, which

help enterprises shift from "empirical decision-

making" to "data-driven decision-making". The

company is alerting enterprises to balance automation

and human intervention to avoid the risk of

technological rigidity amplification. Provide policy

references for the government to support the

intelligence of cross-border logistics infrastructure

and promote the synergy of international rules, so as

to reduce the institutional barriers to the application

of AI technology.

Based on the limitations of the theoretical

research in this study, future research can deepen the

following directions: firstly, to develop sourcing

models with anti-conceptual drift capability to

enhance decision-making robustness under extreme

events, secondly, to optimise the cultural context

comprehension capability of NLP models, and lastly,

to consider adopting empirical research means to

quantify the contribution of AI technology to the

resilience indicators of cross-border e-commerce.

EMITI 2025 - International Conference on Engineering Management, Information Technology and Intelligence

304

REFERENCES

Agca, S., Birge, J. R., Wang, Z. A., & Wu, J. (2023). The

impact of COVID-19 on supply chain credit risk.

Production and Operations Management, 32(12), 4088-

4113

Christopher, M., & Peck, H. (2004). Building the Resilient

Supply Chain. The International Journal of Logistics

Management, 15(2), 1-14

Gao, Y., Feng, Z., & Zhang, S. (2021). Managing supply

chain resilience in the era of VUCA. Frontiers of

Engineering Management, 8(3), 1-6

Liu, F., & Qin, X. T. (2025). How Does U.S. Policy of

Technology Decoupling from China Affect the

Resilience of Enterprise Supply Chains? A Quasi-

Natural Experimental Analysis Based on Difference-in-

Difference Model. Science & Technology Progress and

Policy, 1-10

Liu, S. (2025). Research on the Impact of Artificial

Intelligence Technology on the Supply Chain

Resilience of Automobile Manufacturing Enterprises.

China Business and Market, 1-15

Liu, X. L., & Yang, W. (2024). On the Mechanism of

Influencing Factors of Cross-border E-commerce

Supply Chain Resilience. International Business

Research, 45(03), 16-29

Modgil, S., Singh, R. K., & Hannibal, C. (2022). Artificial

intelligence for supply chain resilience: learning from

Covid-19. International Journal of Logistics

Management, 33(4), 1246-1268

Ponomarov, S. Y., & Holcomb, M. C. (2009).

Understanding the concept of supply chain resilience.

The International Journal of Logistics Management,

20(1), 124-143

Qi, Y. A., Wang, X. R., Zhang, M., & Wang, Q. (2023).

Developing supply chain resilience through integration:

An empirical study on an e-commerce platform. Journal

of Operations Management, 69(3), 477-496

Shankar, V., & Parsana, S. (2022). An overview and

empirical comparison of natural language processing

(NLP) models and an introduction to and empirical

application of autoencoder models in marketing.

Journal of the Academy of Marketing Science, 50(6),

1324-1350

Shen, Z. M., & Sun, Y. Q. (2023). Strengthening supply

chain resilience during COVID-19: A case study of

JD.com. Journal of Operations Management, 69(3),

359-383

The Impact of Artiﬁcial Intelligence Technologies on Cross-Border e-Commerce Supply Chain Resilience

305