The importance of AI-driven efficiency, adaptability, and automation for future

manufacturing

Wei Wu¹, Zhiheng Zhao¹ and George Q. Huang¹

1 Department of Industrial and Systems Engineering, The Hong Kong Polytechnic University, Hong Kong 999077, People’s Republic of China

E-mail: gq.huang@polyu.edu.hk

Status

The future of manufacturing, potentially shaped by Industry 5.0, emphasizes the creation of more human-centric, resilient, and sustainable manufacturing ecosystem capable of mass personalization [1]. Within the transformation, AI is indispensable, fundamentally enhancing efficiency, adaptability and automation across a hierarchy of facilities. Efficiency, in this context, describes the capability of optimizing production processes to maximize output while minimizing resource consumption and operational timelines. AI helps to ensure men, machines, and materials operate cohesively in the right place, at the right time, with minimal inefficiencies [2]. Adaptability denotes the capacity of seamlessly adjusting to dynamic environments, including fluctuating market demands and unforeseen disruptions. The enhanced cyber-physical visibility and traceability empowered by AI facilitate manufacturers to identify disruptions, make data-driven decisions, and quickly adapt processes to meet shifting requirements [3]. Automation concerns the autonomous management and execution of repetitive or complex tasks with minimal human intervention. The convergence of robotics, Internet of Things (IoT), and AI techniques enable individuals or systems to perform accurate, effective, and consistent decision-making, alongside self-learning and self-optimization [4]. In a competitive global market, AI adoption is crucial for maintaining a competitive edge and achieving sustainability objectives. AI technologies are now pervasively deployed in the manufacturing sector [5] (figure 1). Predictive maintenance [6], for example, utilizes AI to analyze sensor data and forecast equipment failures, thereby mitigating downtime and maintenance cost. AI-powered real-time scheduling and execution [7], underpinned by seamless cyber-physical synchronization, enhance production robustness against operational uncertainties and dynamic changes. The automation of manufacturing tasks [8] via AI-driven robotic systems continues to elevate productivity and streamline workflows. Furthermore, AI plays a pivotal role in optimizing supply chains [9] through improved demand forecasting and inventory management. Generative design tools [10] leverage AI to explore extensive design possibilities based on historical prototypes. These diverse applications are converging towards the realization of "smart factories" that ensure highly automated, efficient, and adaptive production environments. Further advances in AI promise even more profound impacts on manufacturing. We can anticipate greater levels of autonomy in manufacturing processes. Enhanced human-machine collaboration will see AI augmenting human capabilities, allowing workers to focus on more complex, creative, and strategic tasks. The ability to offer mass personalization and highly flexible production systems will become increasingly prevalent, allowing manufacturers to respond rapidly to changing market demands and individual customer preferences. The ongoing evolution of generative AI, in particular, is expected to drive further innovation and transformative changes across the manufacturing domain.

Current and future challenges

Despite the unprecedented potential of AI in manufacturing, its broad and effective implementation is impeded by significant challenges. A primary hurdle lies in data-related issues [11]. The efficacy of AI systems is heavily dependent on access to large quantities of high-quality, consistent, and accurately labelled data. However, many manufacturing enterprises grapple with outdated legacy systems, widespread data silos, and a lack of integrated data governance. These limitations often result in datasets that are noisy, incomplete, or poorly contextualized, necessitating laborious and costly pre-processing. Additionally, safeguarding data security and privacy is a paramount concern [12], particularly with the proliferation of distributed AI models. Protecting sensitive manufacturing data and intellectual property from cyber threats remains a critical challenge.

Integration complexity represents another significant obstacle [13]. Modern manufacturing environments are characterized by a heterogeneous technological landscape, wherein advanced information systems coexist with aging legacy equipment that often lacks standardized communication protocols or digital interfaces. The integration of AI solutions into such disparate infrastructures is technically intricate and operationally disruptive. Moreover, the lack of interoperability between AI platforms and off-the-shelf systems, such as Manufacturing Execution Systems (MES) or Enterprise Resource Planning (ERP), further complicates seamless deployment. It requires substantial infrastructure upgrades, which would increase both cost and time. Equally critical are concerns surrounding the safety, reliability, and trustworthiness of AI systems [14]. Ensuring fairness, transparency, and accountability in AI-driven decision-making is important in safety-critical applications. The use of biased or unrepresentative training datasets for AI models can reinforce existing inequities and produce distorted operational outcomes that potentially compromise process efficiency. Furthermore, the opaque nature of many advanced AI algorithms, commonly referred to as the ‘black box’ problem, introduces notable difficulties in validation, debugging, and fostering trust among human operators. Strengthening the robustness of these systems against adversarial attacks and unpredictable variations is essential to maintain long-term reliable operations. Looking ahead, future challenges will prominently feature the need to warrant the scalability and flexibility [15] of AI solutions across diverse and evolving manufacturing environments. Transitioning to industry-wide deployment requires effective AI operations frameworks capable of accommodating increasing data volumes and rising model complexity. Moreover, AI systems must demonstrate enhanced adaptability to shifting production demands, reconfiguring manufacturing cells, or launching new product lines. Addressing these challenges is of paramount significance in unlocking the full potential of AI to revolutionize manufacturing systems on a global scale.

R

Figure 1. A roadmap of AI-driven manufacturing with the assistance of multiple advanced

technologies. The physical layer, at the bottom, illustrates the fundamental elements in manufacturing, man, machine, and material, equipped with IoT devises for data collection, transmission, and computation. The cyber layer is designed to seamlessly mirror physical world through digital twins, while blockchain technology ensures information security and reliability. The top-level application layer leverages various AI techniques based on streams of data to enable a suite of smart services for manufacturing operations.

Advances in science and technology to meet challenges

Several cutting-edge technologies are emerging and evolving to overcome the challenges hindering AI expansion in manufacturing, including digital twin (DT), hierarchical computing, and blockchain (figure 1). These technologies are expected to promote effective integration, robust reliability and scalable deployment. DT technology [16] offers a solid solution for resolving integration complexity and ensuring reliable AI performance. By creating identical replicas of physical assets, processes, and systems in the cyber space, DTs enable real-time monitoring, simulation, and optimization of manufacturing operations. These virtual models can not only facilitate seamless synchronization between legacy systems and AI-driven platforms but augment numerous high-quality data in a generative way, thus considerably elevating model accuracy and consistency. Additionally, DTs enhance transparency and trust by providing a visible sandbox for validating AI algorithms and reducing the latency of decision-making. Cloud-Fog-Edge-End computing architectures [17] establish the scalability and flexibility of AI deployment. Distributing computational resources across cloud, fog, edge, and end devices can enable efficient data processing closer to the source of generation, thereby improving resource utilization and response speed. Such architectures give rise to real-time decision-making in dynamic production environments and also ensure that AI systems can scale across geographically distributed facilities while maintaining operational efficiency. Edge computing, in particular, enhances data privacy and security by processing sensitive information locally, mitigating risks associated with centralized data storage.

Blockchain technology [18] provides a decentralized and immutable framework for enhancing data security, traceability, and trust in manufacturing environments. By enabling secure and tamper-proof data sharing among stakeholders, blockchain alleviates risks of data breaches and ensures compliance with relevant regulations. Smart contracts can automatically enforce access policies and trigger job sequences, tightening the coupling between data provenance and operational control. Furthermore, blockchain can improve the connectivity and traceability of cloud-fog-edge-end computing systems to strengthen AI safety. Finally, advances in generative AI (GAI) [19] and explainable AI (XAI) [20] would reshape the manufacturing by fostering creativity, building trust, and optimizing processes. GAI focuses on creating new and original content based on latent representations learned from historical data. In the industrial context, foundation models trained on multimodal corpora can accelerate domain adaptation and provide human-readable instructions. Meanwhile, XAI aims to make AI system decisions and outputs transparent and understandable to humans. By addressing the Black Box problem inherent in many AI models, XAI facilitates human-AI collaboration while enhancing safety and reliability. Together, these technologies chart a credible path toward powerful, scalable, and trustworthy AI for the manufacturing of the future.

Concluding remarks

In conclusion, the future of manufacturing is deeply intertwined with the advancement of AI. AI serves as a cornerstone for achieving higher levels of efficiency, adaptability, and automation essential to fostering a competitive, sustainable, and human-centric industrial ecosystem. While significant progress has been made in deploying AI within the manufacturing domain, several critical challenges persist, including issues related to data quality, the complexity of system integration, and the imperative for trustworthy and reliable AI systems. Emerging technologies such as digital twins, hierarchical computing, and blockchain present promising avenues to address these challenges by enhancing cyber-physical traceability and visibility, model reliability, process security, and system flexibility. Harnessing these innovations provides the means to overcome existing limitations and fully realize the transformative potential of AI in manufacturing. Moreover, GAI and XAI are anticipated to further accelerate innovation and redefine the manufacturing paradigm. This trajectory will pave the way for smart factories that are not only highly efficient and resilient but also capable of delivering mass personalization, thus securing a competitive edge in the global market and driving the next wave of industrial evolution.

Acknowledgements

This work was supported in part by the Hong Kong RGC TRS Project (T32-707/22-N), in part by Collaborative Research Fund (C7076-22GF), in part by Research Impact Fund (R7036-22), and in part by Innovation and Technology Fund (PRP/007/25LI).

References

[1] Xu X, Lu Y, Vogel-Heuser B and Wang L 2021 Industry 4.0 and Industry 5.0—Inception, conception and perception J. Manuf. Syst. 61 530-535 [2] Guo D, Li M, Lyu Z, Kang K, Wu W, Zhong R Y and Huang G Q 2021 Synchroperation in industry 4.0 manufacturing Int. J. Prod. Econ. 238 108171 [3] Wu W, Zhao Z, Shen L, Kong X T, Guo D, Zhong R Y and Huang G Q 2022 Just Trolley: Implementation of industrial IoT and digital twin-enabled spatial-temporal traceability and visibility for finished goods logistics Adv. Eng. Inform. 52 101571 [4] Sahoo S and Lo C Y 2022 Smart manufacturing powered by recent technological advancements: A review J. Manuf. Syst. 64 236- 250 [5] Nti I K, Adekoya A F, Weyori B A and Nyarko-Boateng O 2022 Applications of artificial intelligence in engineering and manufacturing: a systematic review J. Intell. Manuf. 33 1581-1601 [6] Çınar Z M, Nuhu A A, Zeeshan Q, Korhan O, Asmael M and Safaei B 2020 Machine learning in predictive maintenance towards sustainable smart manufacturing in industry 4.0 Sustainability 12 8211

[7] Ding J, Chen M, Wang T, Zhou J, Fu X and Li K 2023 A survey of AI-enabled dynamic manufacturing scheduling: From directed heuristics to autonomous learning ACM Comput. Surv. 55 1-36 [8] Soori M, Arezoo B and Dastres R 2023 Artificial intelligence, machine learning and deep learning in advanced robotics, a review Cogn. Robot. 3 54-70 [9] Sharma R, Shishodia A, Gunasekaran A, Min H and Munim Z H 2022 The role of artificial intelligence in supply chain management: mapping the territory Int. J. Prod. Res. 60 7527-7550 [10] Mata O, Ponce P, Perez C, Ramirez M, Anthony B, Russel B, Apte P, MacCleery B and Molina A 2025 Digital twin designs with generative AI: crafting a comprehensive framework for manufacturing systems J. Intell. Manuf. 1 1-24 [11] Escobar C A, McGovern M E and Morales-Menendez R 2021 Quality 4.0: a review of big data challenges in manufacturing J. Intell. Manuf. 32 2319-2334 [12] Patel D, Sahu C K and Rai R 2024 Security in modern manufacturing systems: integrating blockchain in artificial intelligence-assisted manufacturing Int. J. Prod. Res. 62 1041-1071 [13] Mao S, Wang B, Tang Y and Qian F 2019 Opportunities and challenges of artificial intelligence for green manufacturing in the process industry Engineering 5 995-1002 [14] Tripathi S, Muhr D, Brunner M, Jodlbauer H, Dehmer M and Emmert-Streib F 2021 Ensuring the robustness and reliability of data-driven knowledge discovery models in production and manufacturing Front. Artif. Intell. 4 576892 [15] Arinez J F, Chang Q, Gao R X, Xu C and Zhang J 2020 Artificial intelligence in advanced manufacturing: Current status and future outlook J. Manuf. Sci. Eng. 142 110804 [16] Wang S, Zhang J, Wang P, Law J, Calinescu R and Mihaylova L 2024 A deep learning-enhanced Digital Twin framework for improving safety and reliability in human–robot collaborative manufacturing Robot. Comput.-Integr. Manuf. 85 102608 [17] Yang C, Wang Y, Lan S, Wang L, Shen W and Huang G Q 2022 Cloud-edge-device collaboration mechanisms of deep learning models for smart robots in mass personalization Robot. Comput.-Integr. Manuf. 77 102351 [18] Leng J, Ye S, Zhou M, Zhao J L, Liu Q, Guo W, Cao W and Fu L 2020 Blockchain-secured smart manufacturing in industry

4.0: A survey IEEE Trans. Syst. Man Cybern. Syst. 51 237-252 [19] Ghobakhloo M, Fathi M, Iranmanesh M, Vilkas M, Grybauskas A and Amran A 2024 Generative artificial intelligence in manufacturing: opportunities for actualizing Industry 5.0 sustainability goals J. Manuf. Technol. Manag. 35 94-121 [20] Alexander Z, Chau D H and Saldaña C 2024 An interrogative survey of explainable AI in manufacturing IEEE Trans. Ind. Inform. 20 7069 - 7081