ON TIME ON TIME ON TIME BUSINESS MANAGEMENT A+ WORK, ON TIME, NO PLAGARIZING; ON TIME

R E V I E W Open Access

© The Author(s) 2025. Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.

Wellbrock et al. Discover Sustainability (2025) 6:886 https://doi.org/10.1007/s43621-025-01808-3

*Correspondence: Wanja Wellbrock [email protected] 1Faculty of Economics, Heilbronn University, Heilbronn, Germany 2Faculty of Management and Sales, Heilbronn University, Heilbronn, Germany

Ethical implications and potential opportunities and risks of artificial intelligence in supply chain management Wanja Wellbrock1*, Margarita Malinovska2 and Daniela Ludin1

1 Introduction The digital transformation is shaping almost all areas of business and challenges com- panies worldwide to adapt to new technologies. One of the key technologies driving this development is artificial intelligence (AI), which optimizes business processes and increases efficiency.

AI in supply chain management offers new opportunities, such as optimizing sup- ply chains, reducing costs, and minimizing the use of materials and energy. However, the widespread influence of AI raises significant ethical questions, as it fundamentally changes people's lives. Therefore, ethical aspects such as data protection, privacy, and security must be considered [1, 2].

This paper examines the practical challenges and ethical aspects of AI applications in supply chain management. Based on different published case studies, it illustrates how AI is applied in practice and which ethical issues must be considered. The research ques- tion is: What practical and ethical challenges arise from the use of AI in supply chain management, and how can companies successfully address them?

This paper does not aim to propose a normative framework or checklist. Rather, it seeks to contribute to the academic discourse by linking established ethical principles

Discover Sustainability

Abstract The digital transformation is reshaping supply chain management through the integration of artificial intelligence. While AI enhances operational efficiency and process optimization, its use raises significant ethical concerns, particularly regarding data privacy, transparency, and bias. This paper examines the practical opportunities and ethical challenges of AI in SCM based on a systematic literature review and the analysis of four industrial case studies. It provides insights into how companies can responsibly implement AI technologies and offers recommendations for promoting transparency, fairness, and security within supply chains.

Keywords Supply chain management, Ethics, Artificial intelligence, Governance, Data privacy

Page 2 of 14Wellbrock et al. Discover Sustainability (2025) 6:886

with real-world use cases of AI in supply chain management. Through this approach, it aims to identify ethically relevant implementation challenges and to highlight areas where theoretical concepts and practical realities diverge.

At the same time, the paper addresses not only researchers in the field of supply chain ethics and technology, but also practitioners, corporate decision-makers, and policy- makers. It aims to support them in understanding both the opportunities and the ethical risks of AI applications in SCM, and in developing appropriate governance measures.

2 Literature review 2.1 Supply chain management

There is no universally accepted definition of supply chain management (SCM). In gen- eral, SCM refers to the optimization of all stages of the value chain, starting from raw material extraction through various production stages to the end consumer. The primary goal is to design the entire process efficiently and cost-effectively while considering cus- tomer requirements.

According to Porter’s value chain approach, SCM focuses on the integration of inter- nal business activities along the entire value chain. This includes aspects such as pro- curement, disposal, recycling, and the social relationships among involved stakeholders. SCM extends the concept of the value chain by integrating internal and external mate- rial, information, and financial flows [3–5].

According to Arndt, the term “Supply Chain Management” can be misleading, as it may create the impression that the primary focus is on supplier-driven supply. Value creation is primarily determined by the end-customer demand. Additionally, the term "chain" suggests a linear structure involving suppliers and customers, whereas SCM con- sists of complex networks with multiple actors and dynamic interactions.

Therefore, the term “Demand Network Management” would be more appropriate. Figure 1, inspired by Arndt’s model, illustrates how SCM—or rather, Demand Network Management—is structured. This visualization highlights the complexity of relation- ships among stakeholders in SCM and how they are interconnected through material, information, and financial flows [6].

Optimizing the supply chain also alters the dynamics of competition. Instead of indi- vidual companies competing against one another, entire supply chains now compete. The market success of a company depends on how effectively it collaborates with its partners [7].

The goal is to maximize customer value without simply shifting costs to suppliers. In this context, SCM increasingly aligns with logistics, but with a stronger focus on cus- tomer-oriented process optimization [8].

2.2 Key challenges in supply chain management and the role of industry 4.0

Some of the central challenges in SCM include increasing product variety, growing demand for product customization, and fragile supply chains. Additionally, there is a ris- ing need for greater transparency across the entire value chain [9].

A crucial approach to addressing these challenges is the concept of Industry 4.0, also known as the “Smart Factory.” This concept refers to the global real-time networking of machines and processes, enabling data exchange and process automation. It focuses on the digital interconnection of manufacturing technologies [10, 11].

Page 3 of 14Wellbrock et al. Discover Sustainability (2025) 6:886

To enhance supply chain connectivity, various technologies such as catalog systems, RFID, and barcode systems are used. Artificial intelligence (AI) plays a crucial role in improving supply capability and demand forecasting by considering influencing factors and optimizing complex tasks such as capacity planning [1, 2].

Additionally, blockchain technology enhances the optimization of material and infor- mation flows by providing an immutable and business-oriented data management system within a decentralized infrastructure [12]. Another recent advancement is the interconnection of physical and non-physical resources, known as the Internet of Things (IoT), which allows machines to interact with each other, leading to more efficient pro- duction processes [13].

Despite these technological advancements, the full integration of supply chains remains one of the biggest challenges, as it requires comprehensive networking and syn- chronization of all stakeholders [14].

2.3 Defining artificial intelligence: concepts and classification

There are various approaches to defining AI. One definition describes AI as a system's ability to perform cognitive tasks that are typically associated with human intelligence.

Fig. 1 Demand network management. Source: Based on [6]

Page 4 of 14Wellbrock et al. Discover Sustainability (2025) 6:886

A system is generally considered intelligent if it can independently and efficiently solve problems. The level of intelligence of such a system is evaluated based on several criteria, including autonomy, complexity, and the efficiency of its problem-solving strategies.

AI is commonly divided into two types: narrow AI, also known as “weak AI,” and gen- eral AI, often referred to as “strong AI.” Narrow AI is designed for specific tasks, such as autonomous driving, chatbots, or voice assistants like Siri. In contrast, general AI, which would possess human-like cognitive abilities across various domains, remains a topic of ongoing research [15, 16].

AI can be divided on different levels. A key subcategory of AI is Machine Learning (ML), a type of narrow AI that enables systems to learn from data and make predic- tions. Unlike traditional software, which operates based on fixed algorithms and remains unchanged after programming, ML can adapt autonomously to changing conditions. Through feedback and continuous learning, ML models refine themselves, allowing for ongoing optimization and adaptability [17].

A more advanced subset of ML is Deep Learning (DL). This technology supports deci- sion-making processes by mimicking the neural structure of the human brain, enabling highly complex pattern recognition and feature extraction [18].

An even more advanced level is Generative AI, which builds upon ML and DL prin- ciples to develop new and creative content. Instead of merely recognizing patterns, gen- erative AI learns from training data and autonomously produces novel results without relying on predefined rules or explicit instructions. Generative AI represents a step toward artificial general intelligence (AGI), which would possess human-like cognitive abilities and transfer problem-solving skills across various domains [19].

2.4 Ethical principles and challenges for artificial intelligence

Ethics is a philosophical discipline concerned with evaluating actions as good or bad. The question of what is "good" or "bad" can only be adequately answered in the context of the circumstances being considered. Ethical evaluation considers the consequences of an action; behaviour is deemed morally acceptable if it does not result in negative consequences.

One of the fundamental principles of ethics was explained by Immanuel Kant: "Act only according to that maxim whereby you can at the same time will that it should become a universal law”. In general, moral values and norms are shaped by cultural, social, and historical factors. This means that assessing potential consequences, particu- larly in emerging and complex fields such as technological development, can be difficult due to a lack of comprehensive knowledge about possible long-term impacts [20].

The ethical foundation for evaluating AI applications is based on principles such as justice, fairness, transparency, and accountability. A fundamental ethical challenge of AI lies in the fact that it is developed for specific tasks and functional roles but lacks human-like complex traits, such as empathy. Unlike humans, AI does not have intrinsic needs or experiences, making it inherently unable to develop a moral conscience [21, 22].

Furthermore, AI is primarily designed to generate profit for companies, giving large tech corporations significant influence over global AI ethics debates. Some critics argue that these companies strategically emphasize ethical guidelines to divert attention from real regulatory efforts, a phenomenon sometimes referred to as "ethics-washing". This

Page 5 of 14Wellbrock et al. Discover Sustainability (2025) 6:886

term describes the practice of promoting ethical principles to enhance corporate image while failing to implement genuine ethical measures. This approach is often used to influence consumer decisions or prevent stricter regulations on AI technologies [23, 24].

In recent years, the ethical debate around AI has increasingly focused not only on indi- vidual moral responsibility but also on institutional and structural accountability. Lead- ing governance frameworks, such as the EU Ethics Guidelines for Trustworthy AI and the OECD AI Principles, emphasize transparency, risk management, and robustness as key elements of responsible AI systems. These frameworks aim to integrate ethical considerations into compliance mechanisms, quality assurance processes, and corporate decision-making structures [25, 26].

Additionally, current discussions highlight that AI ethics must be understood in rela- tion to broader institutional visions of progress. The concept of sociotechnical imaginar- ies highlights how collective expectations about the future shape the way AI is designed and implemented. In supply chain contexts, such imaginaries often focus on efficiency, automation, and optimization, while ethical concerns may be downplayed unless formal governance mechanisms counterbalance these tendencies [27].

Against this background, ethics in AI is not only about normative ideals, but also about organizational practices and policy frameworks. Accordingly, the paper builds on both philosophical principles (fairness, justice, privacy) and applied governance approaches that reflect recent developments in AI regulation, such as the EU AI Act and ISO/IEC 42001. This dual perspective forms the analytical foundation for the case study analysis [22].

2.5 Applications and ethical challenges of AI in supply chain management

A white paper from the Fraunhofer Institute presents expert opinions from various logis- tics sectors on the potential and future perspectives of AI applications in SCM. Based on this, an integrated supply chain model has been developed, synthesizing insights from the ZWF-Advisory-Board, peer-reviewed research, and the Dortmund Application Model for Machine Learning in Logistics. The model illustrates how AI is utilized across different stages of the supply chain (see Fig. 2) [28]:

Fig. 2 Use of AI in the supply chain. Source: Based on [28]

Page 6 of 14Wellbrock et al. Discover Sustainability (2025) 6:886

Procurement and sourcing AI enhances supplier research by analysing publicly available information to assess sup- plier reliability and risk. AI-powered planning algorithms predict customer needs, opti- mizing procurement strategies and reducing supply disruptions. Additionally, AI-driven chatbots improve negotiation efficiency by automating commu- nication, tracking key contractual terms, and generating post-meeting summaries.

Production and manufacturing AI enables predictive maintenance through real-time process and machine data analysis, preventing breakdowns and increasing production efficiency. AI-supported condition monitoring continuously evaluates equipment status, detecting early signs of wear and potential system failures.

Sales and distribution AI-powered algorithms enhance demand forecasting by leveraging machine learning and data-driven models. This improves inventory management, reduces holding costs, and ensures product availability.

Transportation and logistics AI-driven multi-agent systems optimize route planning by dynamically adjusting trans- port schedules based on real-time traffic and demand fluctuations. This enhances vehicle utilization and reduces fuel consumption, leading to more sustainable logistics solutions.

Returns management and reverse logistics Predictive analytics, a subset of machine learning, is employed to anticipate return pat- terns and optimize reverse logistics. By leveraging historical data and statistical models, AI minimizes forecasting errors and mitigates revenue losses due to inefficient returns processing [28].

These applications illustrate the growing integration of AI into operational logistics and production processes. However, most use cases focus strongly on efficiency, speed, and cost-effectiveness, reflecting a techno-economic logic that often marginalizes ethical considerations. As discussed in Sect.  2.4, such priorities are shaped by sociotechnical imaginaries that envision AI as a driver of automation and competitiveness, rather than as a system requiring institutional safeguards.

Beyond all these potentials, SCM faces not only technological and organizational aspects, but also ethical challenges. According to a 2024 survey conducted by Statista among 521 executives in Germany, the key barriers to AI implementation include poor response quality and concerns regarding data privacy and security [29].

One of the central ethical concerns in AI-driven decision-making is transparency and explainability. In supply chain management, additional focus areas include mitigating bias and ensuring data confidentiality. Therefore, the ethical challenges of AI in SCM can be categorized as follows:

2.6 Bias and discrimination in supply chain management

Bias refers to systematic errors or prejudices embedded in datasets. These biases occur when AI systems are trained on datasets that contain pre-existing societal prejudices or incomplete information, leading to distorted or flawed decision-making. In supply chain

Page 7 of 14Wellbrock et al. Discover Sustainability (2025) 6:886

management, biased AI can result in unfair supplier selection, discriminatory pricing, or an exclusion of certain vendor groups. Such risks are often not evident during technical development but emerge during implementation. This underlines the need for account- ability structures that go beyond technical validation and include ethical impact assess- ments and participatory mechanisms. [30, 31].

2.7 Transparency and explainability of AI decisions

A major challenge in AI-driven SCM is the lack of transparency, as AI models operate through complex algorithms that are not always interpretable. This opacity can lead to: Loss of domain expertise, reduced flexibility in decision-making, security risks due to software vulnerabilities and cyberattacks as well as high adaptation costs for existing supply chain infrastructures.

Additionally, research shows that communication gaps, organizational resistance to change, and scepticism towards AI-based decision-making hinder adoption. AI imple- mentation requires shifts in workflows and corporate structures, demanding new skill sets from employees [32, 33].

Here, explainability is not just a technical issue but a governance question. Who is responsible for decisions made by AI systems? Current approaches to AI governance highlight the need for role clarity, traceable system operations, and enforceable legal remedies in cases of harm or malfunction. Without such frameworks, transparency remains a vague aspiration [25, 26].

2.8 Data privacy and confidentiality in the supply chain

AI in SCM relies on the collection and processing of vast amounts of data from different types of economically independent partners, posing significant privacy risks. Challenges arise when data is collected without consent, information is used beyond its intended purpose or sensitive business, or personal data is shared with third parties.

Failure to uphold ethical and legal standards can result in regulatory violations, repu- tational damage, and loss of trust among business partners and suppliers [34, 35].

This tension between data-driven optimization and data protection highlights the need for clearly defined data governance responsibilities. As the EU AI Act and ISO/IEC 42001 propose, privacy-by-design principles, data minimization, and access controls should become standard elements in AI-supported SCM.

3 Methodology The methodology of the paper is based on a systematic literature review, which is an essential part of any scientific work. It is the process of collecting and evaluating relevant information and data on a particular topic. The literature review should reveal the cur- rent state of research.

For this paper, the literature was obtained by searching the university's own research tool BOSS, as well as the World Wide Web, more specifically Google Scholar. While Google Scholar was generally used with keywords relevant to the topic and chapters, the BOSS research tool allowed for a more precise search. In this system, the authors speci- fied keywords, a publication period and that the papers listed should be peer-reviewed. This was the first step in the filtering process of relevant literature.

Page 8 of 14Wellbrock et al. Discover Sustainability (2025) 6:886

To ensure the quality of the papers, the German VHB Jourqual 3 and the Resurchify website were used to check the journals in which the selected papers appeared [36].

Based on the results of the literature review, this study applies a structured ethical analysis to four industrial case studies [37]. These case studies were selected as they rep- resent real-world implementations of AI technologies across different stages of the sup- ply chain, including procurement, production, logistics, and maintenance. They provide detailed and well-documented descriptions of AI applications, enabling a systematic examination of both technical functionalities and associated ethical implications.

The rationale for choosing these specific cases lies in their thematic and functional diver- sity, their practical relevance, and their alignment with the key ethical dimensions iden- tified in the literature: bias and discrimination, transparency and explainability, as well as data privacy and confidentiality. Moreover, the use of peer-reviewed, secondary case studies ensures replicability and methodological transparency, allowing future research to build upon the present framework and apply it to additional empirical contexts.

This combined methodological design – consisting of a structured literature review and the application of ethical categories to secondary case studies – is well supported in the lit- erature. As Snyder (2019) outlines, literature reviews can serve as an independent research methodology that allows for theory development and category building [38]. In addition, the structured use of case studies for exploratory analysis, as described by Yin (2014), enables the contextual grounding of abstract concepts, which is particularly relevant in interdisciplinary research fields such as AI ethics in supply chain management [36].

4 Implementation of artificial intelligence in supply chain management: a case study analysis 4.1 Overview of the case studies

In total, the study covers the following four case studies:

Case study 1: AI in sales and procurement The first company specializes in custom manufacturing, where an AI-driven Enterprise Resource Planning (ERP) system replaces a rule-based system for finalizing offers and product specifications. This constraint satisfaction programming (CSP) is implemented as a web-based interface that provides sales teams and customers with real-time product information. By leveraging shopping basket analysis, product selection patterns are identified, feed- ing into R&D processes. The AI-driven configuration system improves quotation times, document quality, and reduced manual efforts. The system is continuously updated throughout the product lifecycle [37].

Case study 2: AI-driven production planning In the second case study, AI is deployed to assist production planners by automating decision-making. The machines are connected to a cloud-based AI that dynamically adjust production schedules based on real-time system changes and customer orders. The AI uses a genetic algorithm (GA) to optimize production planning, material selec- tion, and tooling changes. The goals are to increase machine utilization, improve pro- duction flexibility, and separate physical resource control from planning processes.

Page 9 of 14Wellbrock et al. Discover Sustainability (2025) 6:886

Additionally, a connected service for customers is developed, allowing them to plan pro- duction online [37].

Case study 3: AI-based quality control in production The third company implements AI-driven camera stream analysis to enhance quality control in food production. The AI system detects defects in packaging in real time and verifies adherence to quality standards.

The objectives are to automate quality control without additional staff, to establish a systematic learning model for defect detection, and to reduce production waste. AI enables more precise quality inspections and accelerates training on various defect types [37].

Case study 4: AI for predictive maintenance in construction equipment In the fourth case, a company integrates Internet of Things (IoT) technology and ML for monitoring construction machinery and improving maintenance processes.

Using time-series data, ML algorithms identify irregularities in machine performance, triggering automatic maintenance requests when a predefined safety threshold was reached. The service portal then automatically schedules maintenance tasks, spare parts orders, and owner notifications. The key objectives are to extend machine lifespan, tran- sition to condition-based maintenance, and reduce operating and lifecycle costs [37].

4.2 Discussion of the case studies

The case studies highlight the diverse applications of AI in SCM, showcasing its impact on efficiency, cost reduction, and sustainability. They demonstrate that AI fundamentally enables the automation of less complex tasks and provides precise recommendations for process improvements. However, ethical challenges such as data privacy, transpar- ency, and algorithmic bias remain critical concerns. Based on the four case studies these aspects are now analysed in detail:

4.2.1 Case study 1: AI in sales and procurement

In the first case study, constraint satisfaction programming (CSP) is used to analyse cus- tomer offers and orders, optimizing the sales process. This application can be used in the supply chain within procurement, purchasing, sales, and distribution. From an ethical perspective, it must be considered that a fundamental prerequisite for processing data is the availability of high-quality data.

According to a study by the Fraunhofer Institute, which surveyed 57 industrial com- panies, only 5% of the companies reported having very good data quality, while only 24% had a good data quality. Considering the three ethical challenges, there is a need for action in terms of bias and discrimination. Poor data quality can lead to distortions and, consequently, to discrimination.

An example is an AI-based recruitment system developed by Amazon, which auto- matically analysed applications. Since the historical data predominantly showed male hires in certain positions, the AI assessed female candidates less favourably.

Regarding transparency and explainability, the transparency of data must be critically examined. All employees must be able to understand how the AI arrives at its decisions.

Page 10 of 14Wellbrock et al. Discover Sustainability (2025) 6:886

The final ethical challenge is data protection and confidentiality. The question arises as to who is responsible when the AI generates incorrect or inappropriate offers. The processing of customer data is particularly sensitive regarding privacy concerns. This highlights the need to improve data quality to enable comprehensive analyses and optimizations. Furthermore, the case illustrates the importance of clearly defined responsibility for data governance. Without defined roles and internal control mechanisms, ethical standards remain vague [37].

4.2.2 Case study 2: AI-driven production planning

In the second case study, the company implements a cloud-based AI with a genetic algo- rithm (GA) to optimize production schedules and improve machine capacity utilization. Within the supply chain, this application is part of the production process. From an ethi- cal perspective, it must be considered that GA is a complex application that is difficult to control.

Due to the large number of considered solutions, there is a risk that decisions are made that are not comprehensible to users, leading to a lack of transparency. A study by the Karlsruhe Institute of Technology demonstrated that unexplained AI decisions strictly reduce the trust in AI systems.

The analysis of the case study shows that from an ethical perspective, ensuring the traceability of decisions is crucial. Otherwise, accountability for mis-planning remains unclear since the GA limits the responsibility of production planners through automated rescheduling.

Concerning bias and discrimination, as well as data protection and confidentiality, there is also a need for action. Processes should be developed to ensure that AI-driven decisions are ethically justifiable, including implementing measures to detect and mini- mize biases in existing data.

Additionally, a focus should be placed on explaining AI decisions in a way that every user can understand. This fosters trust in the technology and allows users to critically evaluate and modify results if necessary.

The missing internal checks and fallback procedures in this case study show that explainability alone is not enough. Clear responsibilities and oversight structures are also needed to ensure ethical use [37].

4.2.3 Case study 3: AI-based quality control in production

In the third case study, camera data is collected from production lines to optimize qual- ity control. This application is categorized under production within the supply chain. Using a deep neural network (DNN), visual inspections are conducted to identify pro- duction defects as well as analyse and resolve machine issues.

Since deep neural networks are used in safety–critical and quality-relevant areas, ethi- cal considerations are essential. Errors in AI can have profound consequences for human safety. Additionally, the use of cameras poses a risk of capturing employees, potentially violating privacy regulations.

A study from the automotive industry also shows that deep neural networks are not yet fully developed, emphasizing the need for additional diligence and further devel- opment. Considering tolerance limits in quality control, the Six Sigma concept is used

Page 11 of 14Wellbrock et al. Discover Sustainability (2025) 6:886

to adjust functional limits based on environmental influences and aging effect. Due to the complexity of neural networks, input data and structures are considered uncertain. Research in adversarial attacks has shown that even minimal changes (e.g., a single pixel) can lead to misclassification, highlighting the vulnerability of neural networks.

From an ethical perspective, regarding explainability and discrimination, AI appli- cations in areas such as driver assistance systems are not yet sufficient, as they cannot guarantee safety. As a result, it is necessary to further develop neural networks to pre- vent data distortions.

This case also shows that the focus on speed and automation can lead to neglect- ing important aspects like control and fairness. Without safeguards like bias checks or human oversight, ethical risks remain part of the system [37].

4.2.4 Case study 4: AI for predictive maintenance in construction equipment

In the last case study, IoT technology and machine learning (ML) are used to monitor and improve maintenance processes. This enabled early detection of wear and reduced operating and lifecycle costs. Additionally, processes such as service orders or spare parts orders were automated. Within the supply chain, these applications are also part of the production process.

From an ethical perspective, it must be considered that there is no control over how data used for IoT and ML is utilized. There is also a discrepancy between data collection and the simultaneous protection of user data. This raises questions about data protec- tion and the resulting confidentiality concerning technologies.

Ethical design is proposed as a solution. A study introduces this concept as a way for users to maintain control over their personal data and information. Normally, these decisions are embedded in algorithms by programmers. Ethical Design allows users to make their own decisions, leading to increased transparency.

At the same time, it remains unclear who is responsible for managing and protecting data across the different supply chain partners. This shows how important it is to make roles and responsibilities part of the technical setup, not just treat ethics as something extra [37].

As a summary, the four case studies prove that AI cannot replace humans but should serve as a supporting tool to relieve specialists of repetitive tasks. Therefore, three main recommendations for companies working together in supply chains can be highlighted:

(1) Data management: To enable the effective deployment of AI systems, companies should routinely evaluate and refine the quality and structure of their data to ensure usability, consistency, and system compatibility.

(2) Employee integration: AI should support employees, not replace them. Therefore, employee training on AI Tools is very crucial for the transparency and explainability of AI decisions.

(3) Transparency and accountability: Companies must establish clear processes, define liability frameworks, and implement ethical guidelines to ensure a responsible use of AI technologies.

These recommendations point to the need for an institutional embedding of ethical principles. Governance frameworks, sociotechnical imaginaries, and clearly defined roles are essential to turn ethical aspects into practical accountability. Building on this

Page 12 of 14Wellbrock et al. Discover Sustainability (2025) 6:886

perspective, organizations should focus on translating these ethical imperatives into concrete practices within their supply chains. Strengthening data governance, enhanc- ing employees’ ability to critically engage with AI systems, and incorporating ethical design elements such as privacy-by-design and bias mitigation into technological infra- structures are crucial steps. Such measures support the alignment of AI adoption with societal expectations and regulatory standards, ensuring trust and accountability across increasingly complex supply chain ecosystems.

5 Conclusion This paper examines the practical and ethical challenges of using AI in SCM. The research question—what practical and ethical challenges arise and how companies can address them—is analysed in detail.

The case studies demonstrate that AI applications can significantly optimize differ- ent areas of the supply chain. At the same time, it became evident that implementing AI comes with ethical challenges. In some application areas, AI is not yet sufficiently mature, bias remains an issue, leading to discriminatory outcomes, and data protection concerns remain unresolved.

These challenges align with recent developments. For instance, the Italian Data Pro- tection Authority temporarily banned ChatGPT due to violations of the General Data Protection Regulation Act (GDPR) [38]. Additionally, the European Commission intro- duced a framework for trustworthy AI, outlining seven key principles:

1. Human agency and oversight 2. Technical robustness and safety 3. Transparency 4. Diversity, non-discrimination, and fairness 5. Societal and environmental well-being 6. Accountability 7. Privacy and data governance [26]

A key discussion is the balance between human decision-making and AI autonomy. AI should support human decisions rather than replace them. Technical and organizational measures are necessary to ensure trustworthiness, including evaluation checklists that help developers and users to assess AI systems [39].

The contribution of this paper lies in the structured application of ethical categories to concrete AI use cases, allowing for a deeper understanding of the ethical implications and practical barriers that arise in SCM contexts. This case-based approach supports the development of more grounded theoretical insights and can serve as a foundation for future empirical or normative work.

The insights derived from the case studies are particularly relevant for supply chain professionals and managers tasked with implementing AI-based systems, as well as poli- cymakers who seek to establish ethical and regulatory frameworks for responsible AI deployment in logistics and operations contexts. By making the ethical challenges visible and relatable, the paper also contributes to a more informed and nuanced public and academic discourse.

AI in SCM presents enormous potential, but it must be approached with caution and responsibility. A careful, ethical approach is essential to maximize benefits while

Page 13 of 14Wellbrock et al. Discover Sustainability (2025) 6:886

minimizing risks. Additionally, deeper research into AI’s decision-making processes is necessary for full transparency and trust.

This study also contributes to the theoretical discourse on ethical AI governance in supply chain management by linking philosophical principles such as fairness, account- ability, and transparency with real-world implementation challenges. By systematically applying these ethical categories to secondary case studies, the analysis highlights how abstract governance frameworks intersect with practical decision-making processes in complex supply chain networks. This case-based perspective extends existing theoreti- cal discussions by emphasizing the institutional and sociotechnical dimensions of ethical AI deployment, offering an understanding of where governance ideals and operational realities converge or diverge.

Future research should explore how AI governance mechanisms—such as role clarity, auditability, or redress options—can be operationalized in SCM contexts. Comparative empirical studies across sectors could help uncover patterns in how ethical risks emerge and are managed in practice. Furthermore, more work is needed on how sociotechnical imaginaries influence AI adoption and shape ethical expectations along supply chains. Such research would support the development of more context-sensitive, responsible AI strategies.

At the same time, the findings of this study offer practical implications for organiza- tions aiming to implement AI responsibly today. Embedding ethical considerations directly into governance structures and operational workflows, establishing transpar- ent decision-making processes, defining accountability frameworks, and empowering employees to critically engage with AI systems are crucial steps. By adopting such mea- sures, companies can balance efficiency gains with the imperative of maintaining trust, fairness, and compliance in complex supply chain ecosystems. Author contribution Conceptualization, WW and MM; methodology, WW and MM; validation, DL; formal analysis, MM; investigation, WW and MM; writing—original draft preparation, MM; writing—review and editing, DL; visualization, MM; supervision, WW and DL; project administration, WW. All authors have read and agreed to the published version of the manuscript.

Funding Open Access funding enabled and organized by Projekt DEAL. No financial funding was received for this study.

Data availability No datasets were generated or analysed during the current study.

Declarations

Ethics approval and consent to participate Not applicable.

Consent for publication Not applicable.

Competing interest The authors declare no competing interests.

Received: 6 May 2025 / Accepted: 14 August 2025

References 1. Min H. Artificial intelligence in supply chain management: theory and applications. Int J Logist Res Appl. 2010;13(1):13–39. 2. Toorajipour R, Sohrabpour V, Nazarpour A, Oghazi P, Fischl M. Artificial intelligence in supply chain management: a system-

atic literature review. J Bus Res. 2021;122:502–17. 3. Burgess K, Singh PJ, Koroglu R. Supply chain management: a structured literature review and implications for future

research. Int J Oper Prod Manag. 2006;26(7):703–29.

Page 14 of 14Wellbrock et al. Discover Sustainability (2025) 6:886

4. Shukla RK, Garg D, Agarwal A. Understanding of supply chain: a literature review. Int J Eng Sci Technol. 2011;3(3):2059–72. 5. Power D. Supply chain management integration and implementation: a literature review. Supply Chain Manag.

2005;10(4):252–63. 6. Arndt H. Supply Chain Management: Optimierung logistischer Prozesse, 8th ed., Wiesbaden; 2021. 7. Antai I, Olson H. Interaction: a new focus for supply chain vs supply chain competition. Int J Phys Distrib Logist Manag.

2013;43(7):511–28. 8. Jääskeläinen A, Heikkilä J. Purchasing and supply management practices in customer value creation. Supply Chain Manag.

2019;24(3):317–33. 9. Jede A, Bensberg F, Klein T. Blockchain-Technologie im Supply Chain Management—Anwendungspotenziale und Kompe-

tenzlücken. HMD Prax Wirtschinform. 2024;61(1):266–83. 10. Abdirad M, Krishnan K. Industry 4.0 in logistics and supply chain management: a systematic literature review. Eng Manag

J. 2021;33(2):187–201. 11. Chauhan C, Singh A. A review of industry 4.0 in supply chain management studies. J Manuf Technol Manage.

2020;31(5):863–86. 12. Queiroz MM, Telles R, Bonilla SH. Blockchain and supply chain management integration: a systematic review of the litera-

ture. Supply Chain Manag Int J. 2019;25(2):241–54. 13. Ben-Daya M, Hassini E, Bahroun Z. Internet of things and supply chain management: a literature review. Int J Prod Res.

2019;57(15/16):4719–42. 14. Zhang C, Gunasekaran A, Wang WYC. A comprehensive model for supply chain integration. Benchmarking.

2015;22(6):1141–57. 15. Babu VS, Banana K. A study on narrow artificial intelligence—an overview. Int J Eng Sci Adv Technol. 2024;24(4):210–9. 16. Butz MV. Towards strong AI. KI Künstliche Intelligenz. 2021;35(1):91–101. 17. Jordan MI, Mitchell TM. Machine learning: trends, perspectives, and prospects. Science. 2015;349(6245):255–60. 18. LeCun Y, Bengio Y, Hinton G. Deep learning. Nature. 2015;521:436–44. 19. Feuerriegel S, Hartmann J, Janiesch C, Zschech P, Generative AI. Bus Inf Syst Eng. 2024;66:111–26. 20. Vieweg SH. KI Für das Gute: Künstliche Intelligenz und Ethik, Wiesbaden, 2023. 21. Siau K, Wang W. Artificial intelligence (AI) ethics: ethics of AI and ethical AI. J Database Manag. 2020;31(2):74–87. 22. Ryan M. In AI we trust: ethics, artificial intelligence, and reliability. Sci Eng Ethics. 2020;26:2749–67. 23. Bietti E. From ethics washing to ethics bashing: a moral philosophy view on tech ethics. J Soc Comput. 2021;2(3):266–83. 24. van Maanen G. AI ethics, ethics washing, and the need to politicize data ethics. Digit Soc. 2022;1:9. 25. Veale M, Zuiderveen Borgesius F. Demystifying the draft EU artificial intelligence act. Comput Law Rev Int.

2021;22(4):97–112. 26. Laux J, Wachter S, Mittelstadt B. Trustworthy artificial intelligence and the European Union AI act: on the conflation of

trustworthiness and acceptability of risk. Regul Gov. 2024;18(1):3–32. 27. Kiran AH. Sociotechnical imaginaries and ethical design of AI systems in organizations. AI & Soc. 2023;38:741–53. 28. Murremhof A, Friedrich M, Witthaut M. Künstliche Intelligenz in der Logistik, Dortmund, 2021. 29. Statista, Umfrage zu Herausforderung bei KI Einsatz, 2024, URL: h t t p s : / / d e . s t a t i s t a . c o m / s t a t i s t i k / d a t e n / s t u d i e / 1 4 7 9 6 3 7 / u m f r a

g e / u m f r a g e - z u - h e r a u s f o r d e r u n g e n - b e i - k i - e i n s a t z /, accessed on April 28, 2025. 30. Sanders NR, Boone T, Ganeshan R, Wood JD. Sustainable supply chains in the age of AI and digitization: research chal-

lenges and opportunities. J Bus Logist. 2019;40(3):229–40. 31. Ferrer X, van Nuenen T, Such JM, Coté M, Criado N. Bias and discrimination in AI: a cross-disciplinary perspective. IEEE Tech-

nol Soc Mag. 2021;40(2):72–80. 32. Balasubramaniam N, Kauppinen M, Rannisto A, Hiekkanen K, Kujala S. Transparency and explainability of AI systems: from

ethical guidelines to requirements. Inf Softw Technol. 2023;159: 107197. 33. Hois J, Theofanou-Fuelbier D, Junk AJ. How to achieve explainability and transparency in human AI interaction. In: HCI

International 2019. Communications in Computer and Information Science. 2019. vol. 1033. 34. Ogbuke NJ, Yusuf YY, Dharma K, Mercangoz BA. Big data supply chain analytics: ethical, privacy and security challenges

posed to business, industries and society. Prod Plan Control. 2020;33(2/3):123–37. 35. Mir AA. Transparency in AI supply chains: addressing ethical dilemmas in data collection and usage. MZ J Artif Intell.

2024;1(2):2–3. 36. Snyder H. Literature review as a research methodology: an overview and guidelines. J Bus Res. 2019;104:333–9. 37. Helo P, Hao Y. Artificial intelligence in operations management and supply chain management: an exploratory case study.

Prod Plan Control. 2022;33(16):1573–90. 38. Bolici F, Varone A, Diana G. Unpopular policies, ineffective bans: lessons learned from ChatGPT prohibition in Italy, ECIS

2024 Proceedings. Impact of AI on Organizations and Society, 2024. vol. 11. 39. Degallier-Rochat S, Kurpicz-Briki M, Endrissat N, Yatsenko O. Human augmentation, not replacement: a research agenda

for AI and robotics in the industry. Front Robot AI. 2022. https://doi.org/10.3389/frobt.2022.997386.

Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

© The Author(s) 2025. This work is published under http://creativecommons.org/licenses/by/4.0/(the “License”). Notwithstanding

the ProQuest Terms and Conditions, you may use this content in accordance with the terms of the License.