Systematic literature review: adoption of artificial intelligence technologies for sustainability in the ready-made garment sector (2013–2024)

The global manufacturing sector is undergoing a profound transformation driven by rapid advances in artificial intelligence (AI), data analytics and digital automation. In manufacturing, AI applications have demonstrated substantial benefits, including predictive maintenance systems that can reduce machine downtime by nearly 50%, automated quality inspection systems achieving accuracy levels above 95% and dynamic demand forecasting systems that significantly reduce inventory waste (Pasteur, 2024). These developments suggest that AI is no longer an experimental technology, but rather a strategic necessity for firms seeking competitiveness, resilience and sustainability.

Despite the global momentum, AI adoption in Bangladesh’s manufacturing sector remains at an early stage. Approximately 12% of manufacturing firms are considering AI applications, while fewer than 5% have implemented AI technologies in any operational capacity (Dong et al., 2020). These limitations are especially consequential for Bangladesh’s Ready-Made Garment (RMG) industry, which represents the backbone of the national economy and accounts for approximately 84% of export earnings (BGMEA, 2022). However, manual production planning, fragmented information flows and weak demand forecasting contribute to an estimated US$2.5bn in annual losses arising from supply chain inefficiencies, forecasting errors and inventory mismanagement (World Bank, 2021). Moreover, 78% of factories lack real-time supply chain visibility, resulting in frequent stockouts and delivery delays that cost an additional US$1.2bn annually (BGMEA, 2022; Ahmed, 2023).

AI holds significant promise for advancing sustainability in the RMG sector. Bangladesh’s textile and garment productions are highly resource-intensive and environmentally polluting. Textile factories consume up to 1,500 liters of water to produce one kilogram of fabric, while dyeing and washing processes discharge toxic effluents into rivers, damaging aquatic ecosystems (Hoque et al., 2022; Sakamoto et al., 2019). Approximately 70% of chemical dyes used in textile operations ultimately enter the environment, and the sector contributes around 15.4% of national greenhouse gas emissions, due largely to coal-based energy use and waste incineration (Islam et al., 2020). AI-enabled sustainability tools – such as digital twins, carbon footprint analytics and optimized production scheduling – offer pathways to reduce water consumption, energy use and material waste. Machine learning-based demand forecasting can minimize overproduction, while AI-assisted fabric utilization planning can reduce cutting waste and excess inventory (Lee and Lim, 2022; Zhu et al., 2022).

Globally, leading apparel producers are increasingly leveraging AI-driven supply chain solutions to address similar challenges. In contrast, Bangladesh’s RMG sector exhibits only limited application of computer vision for quality inspection and basic predictive analytics for planning, with virtually no integration of AI into sustainability-oriented supply chain initiatives (Rafid et al., 2024; Uddin, 2025). AI adoption within Bangladesh’s RMG sector is technically feasible and potentially transformative. AI-powered systems such as the Smart Worker Tracking and Monitoring System (SWTMS), employ computer vision and analytics to improve workplace safety, monitor factory conditions and streamline operations (Hasan et al., 2023; Saha, 2024). In theory, such tools could help Bangladesh’s RMG sector address an estimated 200,000 tons of textile waste annually and reduce overproduction-related losses approaching US$3bn per year (Reza and Du Plessis, 2022; Khan, 2024). However, empirical evidence on how these technologies can be adapted to Bangladesh’s institutional, infrastructural and economic context remains scarce.

The academic literature reflects this imbalance. A systematic review of studies published between 2018 and 2024 reveals that nearly 90% focus on labor conditions and environmental compliance, while only 5% address AI applications, and none directly examine AI’s role in sustainable supply chain management within Bangladesh’s RMG sector (Shuvo et al., 2025). Existing research on social sustainability relies predominantly on small-scale qualitative interviews, limiting generalizability and offering limited insights into technology-enabled interventions (Uddin and Rashid, 2022). Environmental studies largely emphasize quantitative pollution indicators without exploring how digital technologies can optimize resource use (Kabir et al., 2022). Economic analyses tend to focus on macro-level export trends, providing little micro-level understanding of how small and medium-sized enterprises (SMEs) might leverage AI to cope with rising costs and global competition (Rafique and Islam, 2025).

This fragmented body of knowledge underscores a critical research gap at the intersection of AI, sustainability and supply chain management in Bangladesh’s RMG industry. While global scholarship increasingly adopts the Triple Bottom Line (TBL) framework to assess environmental, social and economic performance, few studies operationalize this framework in developing-economy apparel contexts using AI-enabled perspectives (Di Vaio et al., 2020; Stroumpoulis and Kopanaki, 2022). Consequently, policymakers and industry leaders lack evidence-based guidance on how AI technologies can simultaneously enhance competitiveness and sustainability.

In response, this study seeks to examine how AI-driven technologies can be integrated into supply chain management to support environmental, social and economic sustainability in Bangladesh’s RMG sector. By synthesizing theoretical insights from Diffusion of Innovation (DOI), Technology–Organization–Environment (TOE), Resource-Based View (RBV) and Institutional (INT) theories, along with empirical evidence, the research aims to develop a contextualized understanding of AI’s potential, constraints and strategic implications for a developing-country manufacturing ecosystem. In doing so, the study contributes to closing a critical knowledge gap and provides a foundation for designing practical, scalable and sustainable AI adoption pathways for one of the world’s most important apparel-producing nations. Therefore, this study examines the previous literature with systematic literature review (SLR) methods. Section 2 provides a brief review of the literature, followed by Section 3, which outlines the research methodology. Section 4 presents the descriptive and content analysis results and discussion, along with an eight-proposition conceptual framework to guide future research. Section 5 discusses theoretical and managerial implications, limitations and directions for future research, while Section 6 concludes the study.

The integration of AI in Industry 4.0 is crucial for enhancing environmental, social and economic sustainability initiatives. Sustainable production in Industry 4.0 requires efficient inventory management, adaptable production lines, environmentally conscious product design and remanufacturing, automated assembly systems, intelligent storage solutions, self-regulating workstations and thorough product and process traceability. Industry 4.0 gear enhance sustainable manufacturing by reducing waste generation and conserving fixed resources. These features, supported by innovative management and leadership initiatives, empower firms to transition supply networks toward more sustainable operations. Previous research indicates that cloud computing and additive manufacturing enhance sustainability by minimizing material waste and optimizing equipment maintenance efficiency (De Sousa Jabbour et al., 2018; Bag et al., 2021).

AI-driven technologies, including cloud computing, Internet of Things (IoT), Big Data Analytics (BDA) and AI, enhance both horizontal and vertical integration throughout the product life cycle, strengthening sustainable manufacturing systems. AI also facilitates energy transparency and circular economy techniques, such as closed-loop production and lifecycle monitoring, which are crucial for attaining sustainability in Industry 4.0 contexts (De Sousa Jabbour et al., 2018). AI in fashion industry includes a wide range of technologies and methods includes, such as virtual try-on, personalized recommendations, AI-driven design, optimizing the supply chain for clothing, predicting trends and sustainability initiatives (Jung and Suh, 2024).

AI-driven digital twins illustrate these capabilities by simulating supply chain processes to identify carbon hotspots and enabling firms to evaluate emission-reduction strategies virtually prior to real-world execution (Zhu et al., 2022). These approaches have been effectively implemented in European manufacturing and logistics, enhancing warehouse energy efficiency and reducing fuel consumption via route optimization algorithms (Di Vaio et al., 2020). In the Bangladesh RMG sector, the adoption of AI remains nascent, hindered by infrastructure obstacles such as disjointed IoT sensor implementation and insufficiently built data ecosystems (Reza and Du Plessis, 2022). Research highlights that extensive adoption of AI in Bangladesh is contingent upon enhancing data quality and modernizing legacy systems (Jum’a, 2022).

The RMG industry is among the most resource-intensive and environmentally harmful sectors in Bangladesh, with considerable consequences for environmental sustainability. Wastewater produced from dyeing and washing procedures, which often contains hazardous chemicals, is routinely released into rivers, leading to significant water pollution and harm to aquatic ecosystems (Sakamoto et al., 2019). About 70% of chemical dyes utilized in textile processes, exceeding 5.7 × 10⁵ tons of synthetic dyes each year, are discharged directly into the environment, hence intensifying environmental damage (Islam et al., 2020).

Social sustainability is a relatively recent advancement in sustainability literature, as earlier research predominantly focused on environmental factors, resulting in limited in-depth examination of social sustainability (Carter and Rogers, 2008). Some scholars assert that the social dimension is “undervalued,” “underexplored,” “undertheorized” and disregarded, leading to insufficient examination and analysis (Morais and Silvestre, 2018). Social sustainability has been gaining increasing importance alongside environmental concerns among international sourcing agencies and buyers (Huq et al., 2014). Stakeholders are compelling companies and organizations to monitor social issues both within the firm and throughout remote supplier networks (Nakamba et al., 2017). Huq et al. (2014) associated social sustainability with human well-being, encompassing concerns such as human and labor rights, health, safe working conditions, child labor, wage discrimination, rights of association, healthy communities and suitable training facilities.

The RMG sector is experiencing increasing economic challenges due to inflation-driven costs, such as energy, raw materials and wages, which have significantly reduced profit margins. These constraints frequently compel factories to adopt unsustainable practices, including wage reductions and the suspension of annual increments and promotions, particularly for senior management (Mottaleb and Sonobe, 2021). The sector exhibits little market diversity, with around 80% of exports reliant on the European Union and USA markets, hence increasing vulnerability to external shocks (Kabir et al., 2022). Moreover, physical constraints such as port congestion, power deficiencies and inconsistent logistics result in production delays and elevated operational expenses (ADB, 2022). Inadequate regulatory harmonization and bureaucratic inefficiencies diminish the industry’s competitiveness and long-term economic viability (LightCastle Partners, 2024).

The rapid, demand-oriented landscape of Business 4.0 has necessitated the RMG sector to emphasize sustainability alongside enhancing operational efficiency. Escalating material and labor expenses, shortened lead times and swiftly evolving consumer demands have intensified the demand for textile and clothing companies to enhance production processes and supply chains. Regular design modifications prompted by seasonality and fashion trends frequently lead to inventory imbalances, waste production and economic losses. The intricacy of shipping and distributing apparel to satisfy changing customer expectations necessitates meticulous logistics coordination, precise demand forecasting, efficient inventory management, real-time data accessibility and adaptable operational practices to reduce environmental impacts and economic inefficiencies (Ahmad et al., 2020a, 2020b).

The RMG industry is progressively implementing Industry 4.0 technology to address these challenges. Smart manufacturing technologies, underpinned by cloud computing, the IoT, radio-frequency identification (RFID) and BDA, are transforming conventional production paradigms. The integration of business information systems with cyber-physical infrastructure improves transparency and traceability throughout the value chain, allowing RMG enterprises to achieve sustainability goals while preserving global competitiveness. Digitalization enhances the resilience, adaptability and environmental sustainability of manufacturing and management processes, promoting the establishment of efficient garment production ecosystems (Ahmad et al., 2020a, 2020b).

Industry 4.0 technologies, including AI, BDA and the IoT, have significantly enhanced automation, connectivity and decision-making across several sectors (Venkatram and Geetha, 2017). Extensive data analytics enhances organizational intelligence and strategic insight (Sun et al., 2016), while AI-driven data analysis is transforming industries such as healthcare and workforce management (Guo and Chen, 2023). IoT-enabled devices and intelligent networks facilitate data-driven innovation through continuous data collection and real-time system integration (Gubbi et al., 2013). Previous research underscores the necessity of BDA for organizational innovation and efficiency (Wamba and Queiroz, 2020; Dubey et al., 2019; Yadegaridehkordi et al., 2020), while real-time data acquisition and automation via industrial IoT markedly enhance productivity and operational performance (Li et al., 2014).

This study formulates a comprehensive theoretical framework by synthesizing the DOI, TOE framework, INT Theory and RBV to explain how organizations assess, adopt and extract value from emerging technologies. These perspectives are not regarded as independent explanations; instead, they represent complementing mechanisms operating at various stages of the technology adoption process.

The DOI hypothesis explains the preliminary cognitive assessment of technological advancements in businesses. Decision-makers evaluate new technologies based on perceived characteristics, including relative advantage, compatibility, complexity, trialability and observability, as outlined by DOI. These perceptions shape managerial attitudes and affect the perceived attractiveness of adopting new technologies. Nonetheless, although the DOI theory elucidates why managers could acknowledge the potential benefits of innovation, positive perceptions alone do not guarantee organizational adoption, since organizations encounter internal and external restrictions that affect implementation choices.

The TOE framework explains the process by which technology perceptions influence real adoption decisions, thereby addressing this problem. The TOE framework posits that the adoption of technology is contingent upon the interaction of three contextual dimensions: technological readiness, organizational competence and environmental variables. Organizational resources, leadership endorsement and technological infrastructure determine a firm’s ability to implement new technologies, whereas environmental factors such as industry competition, regulatory frameworks and market volatility influence external incentives for adoption (Lin et al., 2020; Tsai et al., 2013). In this context, DOI explains the perceived value of innovation, whereas TOE clarifies the organizational feasibility of adoption.

Institutional Theory further explains this concept by emphasizing the influence of legitimacy demands on organizational adoption behavior. Organizations often adopt new technologies not only for efficiency but also due to coercive pressures from regulatory bodies, mimetic pressures arising from competitors’ behaviors and normative expectations from professional networks and industry standards (DiMaggio and Powell, 1983). These institutional constraints compel enterprises to adhere to prevalent technology methods to sustain legitimacy and mitigate uncertainty in their operating environment. Thus, the adoption of technology may be influenced by both rational efficiency factors and the necessity for institutional compliance.

Although DOI, TOE and Institutional Theory explain the determinants of technology adoption, they offer limited insight into the differing performance outcomes businesses encounter following the deployment of analogous technologies. The RBV mitigates this limitation by emphasizing the internal competencies that allow organizations to get value from technical expenditures. The RBV posits that sustainable competitive advantage is achieved when organizations have resources that are valued, scarce, inimitable and non-substitutable (Barney, 1991). In the realm of digital technologies, such resources encompass sophisticated data management capabilities, analytical expertise, organizational learning and dynamic capabilities that enable enterprises to efficiently incorporate technology into their operational processes (Wamba and Queiroz, 2020). Consequently, whereas institutional and environmental factors may compel numerous organizations to implement analogous technologies, firms exhibit considerable variation in their capacity to translate these technologies into enhanced performance outcomes.

Collectively, these theoretical approaches provide a sequential and multilevel explanation of technology adoption. DOI elucidates the cognitive assessment of technological innovations by managers; TOE contextualizes these evaluations within organizational and environmental factors that influence adoption viability; Institutional Theory explains the legitimacy-driven pressures that encourage adoption decisions; and RBV clarifies how firms utilize internal capabilities to transform technological adoption into enduring competitive advantage. Thus, technology adoption must be perceived not as a single decision but as a dynamic organizational process in which cognitive perceptions, contextual factors, institutional constraints and strategic resources interact to influence both adoption behavior and post-adoption performance.

A set of inclusion and exclusion criteria was used to identify the most relevant literature. Snyder (2019) considers the article’s year of publication, language, category and journal. This study considered only the Web of Science (WoS) for relevant papers (Khatib, 2025), although other platforms such as Google Scholar, EBSCO, IEEE and Scopus are also available (Jia et al., 2020; Kamble et al., 2018). Khatib et al. (2022) observed that WoS is commonly utilized in SLRs due to its provision of reliable bibliographic data, reduction of duplication and facilitation of rigorous citation and co-citation analyses, which collectively bolster the credibility and replicability of review outcomes. They further argue that using a single reputable database such as WoS can improve consistency in search results and reduce interference caused by journals with varying quality standards that may be present in broader databases. This study aims to summarize recognized and influential research in the field; thus, utilizing WoS is consistent with previous high-quality SLRs and ensures that the review is based on extensively reviewed academic literature.

The table below outlines the steps followed in the study. Step 1: Identify relevant documents using keywords; Step 2: Select English-language articles published between 2013 and 2024; Step 3: Exclude book, proceedings, symposium papers, workshop papers and conference articles; Step 4: Select journals indexed in the Social Sciences Citation Index; Step 5: Select publications in business economics, environmental science, ecology, computer science, engineering, transportation, communication, public administration, agriculture, operations research, management science and construction technology (Table 1).

Nonetheless, several papers retained after Step 5 focused on technology adoption in a broad or technical context, failing to investigate adoption intention at the human or organizational level, which is crucial to the aims of this analysis. A considerable number of articles concentrated on sectors unrelated to manufacturing or services (e.g. Health care, finance, smart cities, or exclusively engineering-focused applications), or analyzed system performance, algorithms, or implementation results rather than behavioral or organizational adoption processes. Moreover, several studies failed to establish a clear link to sustainability outcomes, green supply chain management (GSCM), or industry-specific contexts pertinent to the garment sector.

Thus, a thorough full-text review was conducted to evaluate substantive relevance according to several criteria:

• explicit emphasis on the intention to adopt or utilize digital technologies;

• pertinence to manufacturing or service-sector contexts;

• incorporation of sustainability or GSCM perspectives; and

• applicability to, or implications for, the garment industry.

Only studies that fulfilled all of these criteria were retained. The final sample of 48 articles constitutes a focused and thematically consistent body of literature, thereby ensuring analytical rigor and enhancing the internal validity and trustworthiness of the systematic review.

The study particularly examines technological, organizational and environmental aspects influencing AI adoption, excluding an analysis of corporate governance mechanisms such as audit committee characteristics, board structure or the quality of sustainability reporting. The research is based on peer-reviewed academic articles identified through a systematic literature search, rather than on firm-level data or samples selected from sustainability reports. Articles were included or excluded according to established criteria including publication quality, temporal scope, language and thematic relevance to the adoption of artificial intelligence and sustainability within manufacturing contexts. Nonetheless, limiting the data set to papers indexed in the WoS and written in English may result in a certain degree of literature selection bias, as pertinent studies listed in alternative databases or published in other languages may have been omitted. Subsequent systematic reviews may mitigate this issue by integrating various academic databases and multilingual sources to enhance the comprehensiveness of the evidence base.

This research does not estimate empirical econometric models or analyze quantitative data at the firm level. Therefore, statistical metrics often presented in empirical research such as R² values, multicollinearity assessments or fixed-effects estimations are not relevant to the current investigation. The study synthesizes existing literature to identify key determinants affecting the adoption of artificial intelligence technologies and organizes them into a conceptual framework based on the TOE Framework, supplemented by complementary theoretical perspectives such as the DOI theory, Institutional Theory and the RBV. Future empirical research may test the propositions developed in this framework using regression analysis, panel data techniques, or structural equation modeling, thereby providing statistical validation and strengthening the theoretical insights derived from this review.

Endogeneity is a common issue in governance and sustainability research, since the interrelations of organizational structures, strategic practices and performance outcomes may be influenced by reverse causality, simultaneity or omitted variables (Khatib, 2025). The current study does not perform econometric estimations; nevertheless, the systematic review approach addresses this by highlighting works that utilize robust empirical methodologies. During the full-text screening phase, particular attention was given to the methodological techniques utilized in the chosen papers. Many of the reviewed studies employ advanced econometric methods, including fixed-effects models, instrumental variable approaches, generalized method of moments (GMM) and difference-in-differences (DiD) estimations, to address endogeneity bias. In accordance with Khatib’s (2025) methodological guidance, the incorporation of peer-reviewed SSCI journal articles helps ensure that the synthesis literature represents research that typically adheres to recognized methodologies for tackling reverse causality and other endogeneity issues. This approach strengthens the reliability of the conclusions drawn from the systematic review.

An exploration was conducted in the WoS database for articles using titles or keywords related to the adoption or application of AI in the manufacturing sector, resulting in 16,757 articles. The papers were subsequently evaluated using VOS viewer, with a focus on keywords, countries, authors and organizations.

The VOS viewer keyword analysis shows considerable interconnectedness among 1,773 keywords and 18 clusters in AI adoption literature. This density suggests that AI adoption in manufacturing is a mature, interdisciplinary study domain, not technical or social silos. The strong cooccurrence linkages reflect progressive knowledge building rather than fragmented interactions. Engineering-focused AI literature has given way to socio-technical understandings of its acceptability. The emphasis on trust, perceived usability, motivation and management preparedness implies that experts are increasingly viewing AI adoption as an organizational transformation rather than a technological advance. This conceptual strength matches the practical reality of manufacturing environments with resistance, skill gaps and institutional inertia. Methodological sophistication and technical accuracy characterize the second AI technology foundations cluster. Machine learning, reinforcement learning, IoT, cyber-physical systems and Industry 4.0 principles form a solid technological framework for adoption.

Human capital, ethics and social networks are positive developments in the evolution of this field. The literature discusses AI-related social issues, such as worker displacement, reskilling, equity and inclusivity. This trend indicates growing normative maturity, as scholars now examine not only AI adoption contexts but also their broader social consequences. Despite these benefits, bibliometric trends suggest theoretical path dependency. The continued reliance on models such as TAM, UTAUT and TPB may constrain conceptual innovation. These models were designed for individual IT adoption and may not reflect the communal, systemic and power-driven dynamics of AI deployment in manufacturing environments. Labor relations, organizational politics, supply-chain interdependencies and regulatory limitations are often misunderstood.

Technical proficiency and contextual awareness differ. Despite their advanced AI methods, technology clusters often assume perfect data accessibility, consistent infrastructure and digital maturity. Bibliometric relevance does not necessarily imply practical value, especially in industries within developing countries. This danger fosters a tech-centric narrative that may disappoint resource-constrained sectors. Moreover, ethical and human capital clusters appear less prominent compared with the behavioral and technical cores. This imbalance suggests that social and ethical considerations are often overlooked in AI adoption frameworks. Despite little empirical evidence, ethical issues including algorithmic discrimination and disproportionate value adoption are commonly addressed descriptively.

Power dynamics, inequality, governance, sector-specific nuances, as well as longitudinal and post-adoption perspectives and Global South contexts are among the themes that remain overlooked or underdeveloped, that the map gently highlights. Bibliographic results in labor-intensive industries like RMG are insightful and concerning. The socio-technical framework is appropriate, however existing models may not account for job insecurity, low digital literacy, gendered labor dynamics and ethical vulnerabilities within the RMG sector. Greater emphasis should therefore be placed on ethical, institutional and workforce transition challenges, rather than focusing primarily on the technological dimensions. In RMG contexts, AI adoption may depend more on equitable governance, inclusive training and trust in hierarchical labor arrangements than on advanced algorithms (Figure 1).

China, the USA and Western Europe dominate global knowledge production, as evidenced by the VOS viewer coauthorship map. Substantial nodes and dense interconnections signify publication output, funding potential and developed research ecosystems. However, centrality within bibliometric networks does not ensure intellectual dominance. Access to funding, prestigious journals, contemporary infrastructure and the predominance of English-language publications constitute structural advantages. A small number of nations dominate a significant share of the 8,984 and 1,763 international connections, revealing a core–periphery structure in global AI research. Advanced economies tend to dominate collaborative networks; while developing and lower-income countries often function as data providers, experimental settings or application contexts, rather than as agenda-setters.

A total of 107 countries, grouped into ten clusters, reflect formal inclusivity; however, the network structure reveals functional exclusion. Countries such as Malaysia, Turkey, India and South Korea are classified as secondary nodes, not due to their industrial capacity, but because of their relatively weaker integration into leading research networks. Collaborations with core nations are prevalent, as opposed to those with South–South or regionally autonomous research networks. Minor loosely connected nodes representing African and Middle Eastern nations (South Africa, Ghana, Egypt and Saudi Arabia) suggest participation in deeper issues without exerting impact. China, the USA and Western Europe emphasize AI ethics, robotics, predictive analytics and industrial automation, illustrating how leading regions define “significant” AI research. These themes correspond to capital-intensive, high-technology manufacturing and Industry 4.0 frameworks. This imbalance highlights a notable gap between global research priorities and the needs of labor-intensive economies.

RMG enterprises encounter informal labor, cost-effective production, gendered labor forces and precarious employment issues, which are inadequately acknowledged. Bibliometric preeminence results in epistemic superiority, standardizing certain industrial paradigms and standards while sidelining others. Bangladesh’s position as an adopter of AI and Industry 4.0 knowledge, rather than a creator, aligns with the network structure. Bangladesh’s RMG sector holds global significance; yet, there exists a disparity between industrial scale and research competencies. This highlights the risk associated with reliance on technology. Wholesale importation of AI systems, standards and implementation frameworks may result in local sectors relinquishing control over data governance, workforce outcomes and sustained innovation. Bibliometric evidence therefore highlights the need for regulatory, institutional and research reforms in Bangladesh, particularly to strengthen transdisciplinary collaboration and integration into international research networks (Figure 2).

The authorship analysis indicates that although more than 41,000 scholars have contributed to AI-in-manufacturing research, only a relatively small core group forms significant collaboration network. Only 345 authors meet the inclusion criteria, and the largest connected cluster comprises just 86 researchers, suggesting that the field remains highly fragmented. Many scholars publish independently rather than engaging in sustained, interdisciplinary collaboration. The VOS viewer map visually reflects this fragmentation, displaying multiple clusters that represent distinct research communities focused on different aspects of AI adoption in manufacturing.

The visualization also reveals emerging hubs of collaboration centered around key authors who connect smaller groups and help shape the intellectual foundations of the field. These influential contributors play a significant role in guiding theoretical directions, methodological approaches and thematic development. However, the presence of several isolated or weakly connected clusters indicates limited knowledge exchange across subfields. This suggests that research is still split up by theme, institution, or geography. In general, the map shows a field that is growing quickly but is still maturing. It is slowly building stronger collaboration networks that are necessary for a deeper, more integrated knowledge of how AI is being used in manufacturing (Figure 3).

Research on artificial intelligence in manufacturing is transforming enterprises, as illustrated by the VOS viewer network map. The WoS data set included approximately 10,000 organizations; however, only 1,309 formed substantial linkages within the final coauthorship network, appearing as dense clusters. The map indicates that AI-related organizational research is widespread, but not consistently advantageous. Only a limited number of universities and research institutions are engaged in highly collaborative and high-impact research. Clustering patterns indicate that these organizations tend to form communities focused on smart manufacturing, robotics, data-driven production management and Industry 4.0.

Hong Kong Polytechnic University, Curtin University, Huazhong University of Science and Technology and Monash University are among the leading institutions in this field. Their prominent positions suggest regular coauthorship, citation, or collaboration with institutions across Asia, Europe and North America in multi-regional research projects. These universities play a key role in advancing AI-driven manufacturing research, often in collaboration with red, green and blue cluster firms and technology startups. While many nodes demonstrate emerging research interest, they often lack access to advanced research infrastructure and global collaboration networks. Regions such as East Asia, Western Europe and Australia form a strong consortium of technologically advanced economies, whereas developing countries show comparatively limited participation and collaboration. Overall, both graphical and textual analyses point to a dynamic yet uneven research ecosystem. A small number of universities drive innovation, while many others focus on building strategic partnerships, enhancing capacity and engaging in international research collaboration (Figure 4).

The SLR approach shows that 48 articles conformed to the requirements. To examine the categories of technological adoption, 10 articles pertain to AI adoption, 6 to BDA, 6 to the IoT, 4 to RFID, 2 to 3D printing, 5 to cloud computing, 2 to Blockchain and 13 to other forms of adoption (Figure 5).

Adoption theories are primarily divided into two major categories: individual adoption and organizational adoption. Individual adoption theories – such as the TPB (Ajzen, 1991), the TAM (Chatterjee et al., 2021) and UTAUT (Kang et al., 2019) – are used to assess users’ or consumers’ intentions to adopt and use a technology. In contrast, organizational adoption is commonly examined using multiple theoretical frameworks, including DOI, TOE, INT, RBV and, in some cases, TAM (Simões et al., 2020). Among the 48 selected articles, 28 use Diffusion of Innovation (DOI/DIT/IDT) theory, 21 use the TOE framework, 15 apply TAM, 6 use UTAUT, 4 adopt INT, 4 use RBV and 2 reference TRA. Furthermore, 14 studies combine DOI and TOE, while 5 integrate TAM (Figure 6).

Previous studies have identified approximately 71 distinct factors used to measure the intention to adopt artificial intelligence technologies. The 11 most frequently examined factors include compatibility (21), complexity (21), relative advantage (20), top management support (17), perceived cost (15), attitude (14), trialability (10), observability (10), perceived usefulness (10), perceived ease of use (10) and trust (10) (Figure 7).

Previous studies indicate that, among the articles examined, 10 focuses on the application of AI in agile organizations, 7 in manufacturing firms and 5 in SMEs. In addition, 4 studies address the supply chain and logistics sector, 3 focuses on the tourism and hospitality sector and 3 examine the construction industry. On average, two studies investigate the application of AI in sectors such as the cattle industry, banking services, healthcare and the broader service sector. Furthermore, a study has been conducted in a variety of domains, including the silk industry, education sector, smart home technologies, disaster relief operations, the automotive sector, packaging and security systems, the aerospace industry, scientific and technological institutions and the oil and gas industry (Figure 8).

Researchers have employed a variety of methodologies, including surveys, expert interviews, mixed methods, case studies and review-based approaches. However, the majority of the studies (40 out of 48) relied on surveys as their primary research method. Three studies adopted mixed-methods, while two employed case study approaches. In addition, two studies used review methodologies, and one utilized expert interview techniques (Figure 9).

This study does not estimate statistical coefficients or perform regression-based empirical analysis. Instead, its conclusions are derived from a rigorous synthesis of existing literature to identify the key determinants of artificial intelligence adoption. Consequently, the arguments presented in this section represent conceptual relationships that require future empirical validation using quantitative methods such as regression analysis, structural equation modeling or other robustness testing techniques.

Machine learning models and predictive analytics offer significant advantages by enabling dynamic demand forecasting, thereby alleviating key challenges such as overproduction and inventory waste in the ready-made garment sector (Lee and Lim, 2022). Rey et al. (2021) identified a strong relationship between IoT adoption in transportation and logistics and perceived benefits. Similarly, Quetti et al. (2012) demonstrated that the use of RFID in perpendicular supply chains enhances process efficiency and effectiveness while reducing the bullwhip effect. Gunasekaran et al. (2017) highlighted the potential of BDA technology to transform organizational processes. Wamba and Queiroz (2020) examined how perceived advantage influences blockchain adoption within the TAM theory. In exploratory research on IoT adoption in logistics and supply chain management, Tu (2018) found that perceived benefits significantly influence firms’ adoption decisions. Likewise, Lai et al. (2018) identified perceived benefits (PB) as an independent variable within technological components to explain how logistics and supply chain management utilize BDA under TOE theory. Their findings indicate that perceived benefits strongly influence adoption behavior. Furthermore, Managers play a crucial part in the company’s readiness to try BDA once they understand its benefits. Based on this understanding, the following proposition is proposed:

The adoption of technologies is often influenced by cost considerations. The economic benefits of AI – including cost reductions, enhanced efficiency and competitive advantage – are well documented in international supply chain studies (Di Vaio et al., 2020). For example, AI-based inventory management systems can reduce overstocking and spoilage, potentially increasing profit margins by 10–30% in the retail sector (Lee and Lim, 2022). In the RMG industry, such techniques could help mitigate the estimated annual loss of USD 3 bn caused by overproduction and unsold inventory in Bangladesh (Reza and Du Plessis, 2022). Mabad et al. (2021) found that RFID significantly reduces operational costs. Effective operations affect costs and therefore cost-cutting innovations are frequently accepted. RFID can cut IT costs dramatically. Cost efficiency plays a critical role in shaping technology adoption decisions, which is based on TOE framework; for example, cost reductions from RFID adopt this enhanced setting (Mabad et al., 2021). Similarly, Lin et al. (2020) argue that high costs can hinder the adoption of GSCM practices, particularly in SMEs. In contrast, Oliveira et al. (2014) demonstrate that cloud computing can reduce infrastructure, energy and maintenance costs, while also shortening system update cycles. These cost savings contribute to enhanced competitive advantage and increase the likelihood of technology adoption (Oliveira et al., 2014). Based on these insights, the following proposition is proposed:

Oliveira et al. (2014) described a security breach as the loss of sensitive or private data by an organization or government entity. Concerns regarding such breaches have intensified in recent years. Satar et al. (2018) warned that security threats associated with IoT adoption may be greater than initially anticipated. The lack of robust industrial data ecosystems and cybersecurity capabilities – particularly among RMG SMEs, where fewer than 20% adopt adequate security frameworks – poses a significant barrier to advanced digital transformation (Narashimman et al., 2024). Mabad et al. (2021) identified RFID data security as a critical issue in technology adoption, emphasizing that security and privacy issues are key components of the technological dimension within the TOE framework. Data privacy concerns can discourage organizations from adopting technologies such as RFID. However, advanced RFID systems incorporating mutual authentication, secure key exchange and encrypted data storage can enhance privacy protection. Similarly, while AI technologies offer substantial benefits, they also raise significant security and privacy challenges, requiring organizations to proactively identify risks and implement appropriate safeguards. However, neither the full sample nor industry-specific subsamples demonstrated that security concerns impeded cloud computing uptake. This may be due to recent privacy-enhancing, surveillance and encryption developments for cloud data security, confidentiality and integrity. FedRamp and the Information Security Management Act have increased cloud-based data security and organizational confidence. This may explain why cloud strategy assessments neglected security and privacy (Oliveira et al., 2014). Overall, it is suggested that:

Lin et al. (2020) observed that highly skilled and educated employees are more likely to adopt and effectively use new technologies. Organizations with strong innovative capabilities are also better positioned to implement advanced environmental practices. In contrast, firms lacking absorptive capacity face difficulties in acquiring and transferring knowledge. Comprehensive employee training can help overcome sustainability-related knowledge barriers. In this context, effective human resource development plays a crucial role in enabling Malaysian SMEs to adopt GSCM. Alsetoohy et al. (2019) argued that workforce quality and resource availability significantly influence technology adoption, consistent with the organizational dimension of the TOE framework. The successful implementation of business IT systems depends heavily on employees’ skills, knowledge and attitudes. Without sufficient technical expertise and confidence, employees are less likely to accept and adopt new technologies (Johnson and Diman, 2017). Positive attitudes toward technology adoption are closely associated with higher levels of creativity and competence. The literature further emphasizes the importance of employee knowledge, skills, experience and organizational investment (Lin, 2018). Employees must understand e-SCM systems and adapt to dynamic environmental conditions in complex supply networks. Effective e-SCM implementation requires strong coordination across supplier networks and collaboration among partners. In an increasingly competitive global environment, e-SCM has become essential for maintaining competitiveness, necessitating sustained investment in human capital. Training and incentive programs are therefore critical organizational investments. The availability of skilled personnel and organizational resources significantly influences a firm’s technological capability, alignment and competitiveness. Consequently, the adoption of AI in Bangladesh’s manufacturing sector – particularly within the RMG industry – is a complex process. It requires not only technological investment but also policy alignment, ecosystem development, industry–academia collaboration and a strong focus on human capital development (M and Chattu, 2021). Furthermore, fostering innovation while maintaining workforce stability necessitates a balanced approach that integrates human expertise with the computational capabilities of AI (Roy, 2024). It is therefore possible that:

Oliveira et al. (2014) described the technological context as the set of an organization’s technological dimensions that enable the adoption of new innovations. Similarly, The TOE framework includes technical context as an organization’s adaptability of innovative compatibility, accessibility and complexity (Mabad et al., 2021). Technological resilience, encompassing IT infrastructure, internet capabilities and qualified individuals, is often seen as a key facilitator of the adoption of new technologies like cloud computing, AI, BDA and IoT-driven GSCM systems. According to surveys, companies with strong IT infrastructure and technical expertise are better positioned to incorporate cloud-based solutions successfully. Similarly, in the Australian silk industry, technical capabilities and farmer self-efficacy have been linked to RFID adoption (Quetti et al., 2012). However, empirical evidence on the role of technical preparedness is conflicting. While some research shows a beneficial impact, others imply that it may have little effect on adoption decisions. For example, Wu et al. (2013) discovered that companies with superior internal information-processing skills may be less likely to adopt cloud computing since their existing systems already satisfy their demands. Therefore, one might argue that:

Competitive pressure denotes the external market dynamics and associated strategic necessities that drive companies to implement advanced technology to maintain or improve their market standing. In dynamic and highly competitive sectors, firms encounter continual pressure to enhance their technological capabilities, not merely to survive but also to attain competitive advantages. Oliveira et al. (2014) emphasized that competitive pressure expedites technology adoption, as companies pursue efficiency enhancements, increased visibility and real-time operational precision. This impetus propels the integration of breakthrough technologies, including cloud computing, RFID, e-commerce, blockchain and various digital tools across commercial sectors.

Likewise, competition also heightens the demand for analytical expertise. Competitive pressure (Oliveira et al., 2014) being explained by institutional theory, highlights external influences like industry norms (Correia Simões et al., 2020a) and pursuit of competitive advantage is influenced by RBV theory (Lutfi et al., 2023). Research conducted by Sun et al. (2020) and Iranmanesh et al. (2023) highlights that intense industrial competition compels enterprises, particularly SMEs, to invest in BDA and associated AI technologies. Organizations that neglect to embrace new technology jeopardize their operational efficiency and strategic market standing.

In this situation, competitive pressure is directly linked to competitive advantage. Companies view sophisticated technologies like AI, machine learning and deep learning as tools for augmenting their competitive advantage through improved decision-making, increased productivity and superior sustainability outcomes. Chatterjee et al. (2021) contended that the quest for competitive advantage enhances the perceived utility and user-friendliness of AI technologies, hence promoting their adoption in manufacturing and production enterprises.

According to the discussion above, in the RMG sector, both domestic and worldwide rivalry compels enterprises to use AI-driven sustainability solutions to satisfy customer requests, adhere to global standards, mitigate environmental consequences and maintain competitiveness in supply chains. Consequently, competitive pressure not only compels enterprises to innovate but also influences the perceived relative advantage of implementing AI technology for sustainability. Consequently, it can be posited that:

Lai et al. (2018) examined supply chain networks to assess drivers for adopting data analytics for logistics and supply chain management. Supply chain information sharing assesses how often a company provides relevant, accurate, comprehensive and confidential data with partners. Goal congruence aids successful supply chain communication. Supply chain collaboration and organizational capabilities require information flow while supply chain (SC) partner connectivity allows information exchange among stakeholders. Gunasekaran et al. (2017) found networking and knowledge exchange increase talent management commitment. Accordingly, SC members require connectivity to communicate and share data. The RMG sectors of Bangladesh is part of the global apparel supply chain. The institutional theory further assesses the pressure from the external environment and a company may adopt a new invention to keep up with its commercial partners while also maintaining internal equilibrium (Correia Simões et al., 2020b). Information sharing and networking boost supply chain visibility, efficiency and customer satisfaction (Lai et al., 2018). A supply chain with effective information sharing and IT integration uses BDA technology more. Therefore, supply chain connectivity via BDA may be beneficial and suitable. Given the aforementioned, the following proposition is made:

Businesses seek to adopt technologies with more economic benefits and performance than competitors. Sustainable practices can reduce expenses associated with industrial waste disposal and treatment, energy efficiency and natural resources, as well as financial, environmental and productivity impacts (Lin et al., 2020). The degree to which an innovation is better than its predecessor is called relative advantage. If current procedures and practices outweigh the technology’s limitations, adoption increases. This context discusses endogenous and exogenous technology adoption components, including Relative Advantages. The DOI theory mention five components which are responsible for organizational adoption of technology, relative advantages is one of the most influential among them (Lin et al., 2020). A company’s decision to accept an invention or technology depends on its superiority and benefits. Merchants’ willingness to employ BDA depends on its perceived advantages over rival technology (Lutfi et al., 2023). Certain textile companies in Bangladesh have commenced the adoption of AI-driven quality control systems. These systems employ computer vision technology and machine learning algorithms to automatically examine clothing for faults with enhanced speed and precision compared to human inspectors. Through the examination of high-resolution photographs of apparel, AI algorithms identify even minor faults and irregularities in stitching, fabric texture and color. Most scholars studying organizational-level technology adoption believed relative advantage was crucial to technology selection. AI, Big Data, IoT, cloud computing, RFID, GSCM and halal meat supply chain adoption are analyzed using relative advantage (Na et al., 2022; Ullah et al., 2021). Thus, the following proposition is proposed:

DOI, TOE model, IT and the RBV are integrated and extended to develop a comprehensive model for understanding RMG industry adoption of artificial intelligence technologies for sustainability. This study synthesizes these theories to assess AI adoption in a competitive, ecologically sensitive and technologically advancing industry like garment manufacturing. Integration raises understanding of internal and external limitations, resource capabilities and strategic motivations that affect emerging technology uptake in current economies. DOI theory critique and extension are this research’s key theoretical contributions. DOI prioritizes innovation and organizational preparation but ignores the external business environment’s impact on adoption. This study fills that gap by combining DOI with the TOE framework to include internal (organizational size, structure and technological capabilities) and external (market rivalry and regulatory pressures) factors. This theoretical integration provides a holistic account of adoption behavior, especially in volatile, demand-sensitive businesses like the RMG industry, where rapid technological adoption is essential for achieving and sustaining competitive advantage.

Institutional theory explains how social norms, industry standards and stakeholder expectations regulate business behavior coercively, normatively and mimetically. These institutional variables are important for RMGs due to global supply chain dynamics, environmental control and customer demands. Social variables, environment and institutional perspectives influence technology adoption decisions. This strengthens the AI adaptability concept in international industries. Organizations integrate AI for competitive resilience because sophisticated IT infrastructure, experienced human resources and robust data capabilities are unique and nonreplaceable assets. This connects the RBV to manufacturing sustainability and digital innovation.

A novel theoretical framework is proposed to reveal previously unexplored linkages between perceived benefit, cost savings, compatibility, security, technology readiness and supply chain connectedness. The research addresses gaps in the literature and contextualizes them to create an original approach for using artificial intelligence technology to enhance sustainability in the Bangladeshi RMG industry. This provides a context-specific, multifaceted, empirically informed methodology for technological adoption to enhance sustainability-focused technology research.

This study’s findings present practical implications for improving environmental, social and economic sustainability in the Bangladesh RMG sector via the implementation of Industry 4.0 technologies. Bangladeshi RMG industries can utilize IoT-enabled water and energy monitoring systems to mitigate excessive resource usage, especially in dyeing and finishing processes. Blockchain-enabled traceability solutions can improve transparency in wage disbursements and working hours, tackling persistent issues associated with labor abuse and wage suppression in the Bangladeshi RMG sector [Amin, Baldacci and Kerbache (2025); Jum’a, 2022].

Industry 4.0 technologies, including AI-based demand forecasting, RFID-integrated inventory management and digital twins, can assist Bangladeshi RMG enterprises in decreasing lead times, minimizing overproduction and enhancing supply chain resilience. These are especially pertinent due to Bangladesh’s reliance on the EU and US demand-markets and its susceptibility to demand variations (Kabir et al., 2022). Moreover, cloud-based business solutions can assist small and medium-sized RMG firms by reducing initial IT investment expenses while enhancing coordination with suppliers and buyers. The contextual implementation of Industry 4.0 technologies in Bangladesh’s RMG industry can bolster sustainable performance and improve global competitiveness. Nonetheless, actualizing these advantages necessitates investments in digital infrastructure, workforce and skill enhancement, data quality augmentation and regulatory assistance to promote technology use throughout the business (ADB, 2022).

Global apparel firms exemplify these advantages: H&M reduced downtime by 25% with IoT-driven predictive maintenance, Zara decreased energy expenses by 20% through smart metering and Nike reduced water usage by 30% through AI-based monitoring. These examples demonstrate how sustainable technologies can produce both economic and environmental benefits (Abidov, 2024). Bekkari and Zeddouri (2019) employed an artificial neural network (ANN) model to regulate and forecast the effluent chemical oxygen demand and efficiency at the Doha West wastewater treatment plant (Nourani et al., 2021).

The policy implications are especially relevant for growing economies like Bangladesh, where the RMG sector encounters ongoing regulatory fragmentation, compliance demands from international buyers and constrained technological capabilities. Policymakers should leverage these findings to provide favorable conditions for AI-driven sustainability by eliminating legal obstacles, implementing exclusive financial incentives and enhancing interorganizational collaboration and knowledge transfer. Training and capacity-building activities are crucial to equip organizations with the necessary skills to apply AI solutions and promote sustainable human capital (Sima et al., 2020).

In Bangladesh, regulatory and institutional obstacles are linked to adherence to several international safety and social standards. Initiatives like the Accord on Fire and Building Safety in Bangladesh and the Alliance for Bangladesh Worker Safety have markedly enhanced fire, electrical and structural safety via regular inspections in accordance with the Bangladesh National Building Code. Global compliance frameworks, including Worldwide Responsible Accredited Production, Business Social Compliance Initiative, Ethical Trading Initiative, Social Accountability International and Sedex (supplier ethical data exchange), enforce stringent labor, environmental and ethical standards on manufacturers in Bangladesh. Nonetheless, traversing concurrent compliance frameworks increases operational expenses and administrative complexities for RMG enterprises. Strategic national policies that incorporate AI deployment within current regulatory frameworks can optimize compliance monitoring, enhance transparency and refine data-driven reporting. Harmonizing AI-driven sustainability initiatives with the demands of prominent EU and North American purchasers would enhance environmentally resilient supply chains and bolster worldwide competitiveness.

To promote sustainable textile manufacturing in emerging markets, better connections between different sectors are needed. Governments, research institutions and industry groups should support collaborative R&D platforms, biotechnology incubation centers and knowledge-transfer initiatives that are specifically designed for textile processing. Combining biotechnology with digital monitoring systems, policies for cleaner production and initiatives for a circular economy would make environmental performance and global competitiveness even better (Rahman et al., 2019).

Facilitating AI adoption in Bangladesh can expedite circular business models, augment resource efficiency and bolster resilience amid the Fourth Industrial Revolution. Coherent policies linking technological advancement with international sustainability standards will be essential for sustaining the long-term survival of the Bangladeshi RMG sector and preserving buyer trust. This report offers practical guidance for managers and policymakers to enhance AI-driven sustainability in the RMG sector and advance circular economy goals.

Alongside technology and physical investments, administrative structures are essential in enabling AI-driven sustainable transformation. Regulators in emerging economies should enhance corporate governance frameworks to ensure firms have the required capabilities to assess and manage digital innovation. A pragmatic strategy is promoting the participation of professionals with financial and technological expertise in audit committees, since they are better equipped to evaluate the financial viability, risk exposure and value of AI investments. This expertise is crucial when companies implement advanced technology necessitating substantial capital investment, cybersecurity measures and data governance systems.

Moreover, legislators and industry associations ought to advocate for sustainability-focused training initiatives for board members, senior executives and supply chain managers. These programs can improve organizational understanding of the role of AI technology in environmental monitoring, resource efficiency and social compliance in global supply chains. Strengthening management competencies via focused training programs will allow companies to effectively link digital transformation efforts with global sustainability standards and consumer expectations. In the Bangladeshi RMG sector, collaborative initiatives among government entities, industry associations and academic institutions might create certification programs and executive training platforms centered on AI-driven sustainable manufacturing to increase corporate governance capabilities and facilitate the transition to technologically advanced and sustainable industrial systems.

This study could be improved in several aspects. The research exclusively examined publications indexed in the WoS database. This indicates that it may have overlooked significant studies accessible in other reputable databases. Secondly, only articles in the English language were taken into account, potentially overlooking significant contributions published in other languages. This study concentrated exclusively on the RMG industry, hence constraining the applicability of the findings to other industries. The study employed exclusively the SLR approach, which, while comprehensive, may not capture the entirety of empirical evidence or theoretical developments available through alternative research methods. The review exclusively examined articles from the past decade. This may have excluded prior efforts that could have provided historical context or long-term insights.

Subsequent studies may in addition examine potential moderating or interaction effects among the determinants mentioned in this review. Organizational attributes, including business size, managerial competence, digital maturity and institutional pressures, may influence the correlation between perceived technological advantages and actual adoption choices. Empirical studies utilizing techniques like structural equation modeling or regression-based interaction analysis could yield deeper insights into the impact of contextual factors on AI adoption outcomes in the RMG sector.

Empirical validation using extensive surveys, longitudinal research, or structural equation modeling would ascertain the robustness, reliability and causal linkages among the identified variables. Subsequent research should expand the literature evaluation by integrating additional academic databases and non-English publications to mitigate linguistic bias and improve the comprehensiveness of evidence base. Broadening the sampling size to encompass all apparel makers, not solely those affiliated with the Bangladesh Garment Manufacturers and Exporters Association, would enhance the generalizability and external validity of the results. Furthermore, qualitative and mixed-method approaches, including interviews and case studies, may yield insights into managerial perspectives, organizational preparedness and contextual issues affecting AI-driven sustainability. Future models should integrate further TOE variables, as well as cultural elements and varied organizational contexts, to more accurately reflect the intricacies of enduring AI adoption. Ultimately, bilingual and cross-cultural research collaborations are advocated to guarantee more inclusive and globally relevant interpretations of AI’s contribution to enhancing industrial sustainability.

This research presents a unique and comprehensive perspective on the incorporation of AI technology for sustainability in the RMG sector. The research indicates that the decision to implement AI within organizations is influenced by technological, organizational and environmental factors than by behavioral theories alone, as evidenced by seventy-one drivers identified in prior literature on the adoption of IoT, BDA, cloud computing RFID and 3D printing across diverse sectors. This research, based on the DOI and TOE frameworks and enhanced by IT and the RBV, provides a comprehensive examination of the factors affecting the potential and difficulties related to AI adoption in this area. Although the prevailing literature emphasizes possible advantages and obstacles, limited research examines the strategic risks linked to the failure to adopt AI, including the loss of competitive position in the global market.

The results indicate that effective AI implementation in the RMG sector depends on perceived relative advantages. These encompass expected advantages such as financial savings, enhanced security and privacy, greater supply chain connectedness and access to proficient people and supportive infrastructure. Competitive challenges and alignment with sustainability goals further underscore the necessity to adopt AI technologies. This study underlines that the integration of AI for sustainability in the RMG industry is not merely a technological progression; it is a strategic need. It necessitates collaborative efforts from enterprises, legislators and worldwide participants to cultivate industrial ecosystems that are robust, ecologically sustainable and internationally competitive. This research provides timely and actionable insights for addressing the interconnected concerns of economic growth and environmental responsibility amid the context of digital revolution and sustainability that characterize contemporary trade in the Fourth Industrial Revolution.