AI enabled manufacturing for new quality productive forces

By the MI Division

In the age of digital transformation, manufacturing is undergoing a profound evolution. Artificial Intelligence (AI) is no longer a futuristic concept—it is a present-day catalyst reshaping how products are designed, produced, and delivered. For Hong Kong, a city historically known for its agile economy and global connectivity, AI presents a unique opportunity to revitalise its industrial base and cultivate New Quality Productive Forces (NQPFs).

NQPFs represent a shift from traditional, labourintensive production to intelligent, data-driven, and high-quality manufacturing. They are essential for achieving New Industrialisation, a strategic direction that aligns with Hong Kong’s broader economic goals. By integrating AI into manufacturing, Hong Kong can enhance productivity, ensure quality, and foster innovation—laying the foundation for a resilient and competitive industrial future.

Hong Kong’s manufacturing landscape

Evolution of manufacturing in Hong Kong

With the rise of Industry 4.0, Hong Kong is witnessing a resurgence in advanced manufacturing, particularly in sectors such as biomedical devices, electronics, medicine, food processing and smart logistics. The key enabler of this transformation is data— the lifeblood of AI systems. As factories digitise their operations, they generate vast amounts of data that can be harnessed to train AI models, optimise processes, and predict outcomes. These transformations have laid the groundwork for a new era of intelligent manufacturing, where AI plays a central role in driving innovation and shaping the future.

Figure 1: Concept of AI enabled smart manufacturing line (AI-generated)

New Quality Productive Forces: Shaping Future Manufacturing with AI

The concept of New Quality Productive Forces emphasises innovation, high-end technology, smartification, sustainability and future skills. AI enables these forces, for example, by:

• Accelerating innovation by enabling rapid prototyping and data-driven design
• Powering high-end technology through intelligent control of advanced robotic system
• Driving smartification by automating and connecting processes for intelligent operations
• Enhancing sustainability by optimising energy use and minimising waste
• Fostering future skills through human-AI collaboration

In this context, AI is not merely a tool—it is a strategic asset that empowers manufacturers to compete globally while aligning with Environmental, Social and Governance (ESG) goals.

Figure 2: Concept of Key AI Enabled Technologies for Smart Manufacturing (AI-generated)

AI revolutionising manufacturing

AI is not replacing engineers— it’s empowering them.

The development of AI in manufacturing

Artificial Intelligence (AI) has evolved from a support tool into a transformative force in manufacturing. In the early 2010s, manufacturers began integrating data collection and full digitalisation through Industrial Internet of Things (IIoT) to enhance automation. These technologies are now applied across every stage of the manufacturing lifecycle—from design and prototyping to production and logistics, enabling manufacturers meeting the 1i and 2i stage of Industry 4.0.

Today, AI systems are capable of learning, reasoning, and adapting in real time. These innovations enable manufacturers to simulate production environments, detect defects with precision, and optimise operations dynamically.

Key technologies include:

• Machine Learning (ML): Enables systems to learn from data and improve over time.
• Computer vision: Allows machines to interpret visual information for quality control and defect detection.
• Natural Language Processing (NLP): Facilitates humanmachine interaction and documentation.
• Digital twins: Virtual replicas of physical systems used for simulation and optimisation.

To advance to the higher maturity levels of 3i and 4i, AI plays a pivotal role. At the 3i level, AI empowers the integration of cyberphysical systems through the Industrial Metaverse, facilitating decentralised decision-making and enhancing human-machine collaboration. Transitioning to 4i, AI further drives intelligent, autonomous processes and self-organising systems, enabling self-optimising operations and autonomous control across the value chain. These advancements are essential for achieving seamless horizontal integration and fostering a smart, efficient manufacturing environment.

Figure 3: Industry 4.0 Maturity Level (Courtesy of HKPC and Fraunhofer IPT)

Recent advancements of AI in manufacturing

A key advancement is the emergence of AI agents, which are categorised into:

• Virtual AI agents — software-based systems that analyse data and support decision-making.
• Embodied AI agents — physical systems such as robots that interact with the real world.

These agents are central to the evolution of smart factories, where AI not only automates tasks but also collaborates with human workers to enhance productivity and safety. (WEF, 2025)

Manufacturing companies are increasingly integrating both virtual and embodied AI agents into their operations. While these technologies are still in the early stages of development and require further refinement for large-scale deployment, it is crucial for industry leaders to proactively address the foundational requirements that will support their effective implementation and seamless integration into existing systems.

Figure 4: Four Types of Virtual AI Agent (Based on Boston Consulting Group and World Economic Forum)

Figure 5: Types of Embodied AI Agents (Adapted from Boston Consulting Group, as cited in World Economic Forum)

AI trends and policies in Mainland China

Mainland China continues to lead globally in AI adoption, particularly in manufacturing. The government’s strategic plans, such as the New Generation Artificial Intelligence Development Plan, emphasise AI as a core driver of industrial modernisation. Chinese manufacturers are rapidly deploying AI for predictive maintenance, supply chain optimisation, and intelligent robotics.

China’s policy environment supports this growth through substantial investments in AI infrastructure, including national AI innovation centres and smart manufacturing pilot zones. The integration of Embodied AI, such as autonomous mobile robots and AI-powered inspection systems, is accelerating across sectors from electronics to automotive. These developments are underpinned by a strong emphasis on data sovereignty, cybersecurity, and AI ethics, aligning with global standards while maintaining national priorities.

AI trends and policy in Hong Kong

Hong Kong is positioning itself as a regional leader in AI innovation through a balanced approach that promotes technological advancement while safeguarding public interest. In April 2025, the Digital Policy Office (DPO) released the Generative AI Technical & Application Guideline, which outlines ethical and operational standards for deploying AI across sectors. This guideline emphasises transparency, accountability, and data security, aiming to build a trustworthy AI ecosystem.

To support this vision, the government has launched several initiatives, including the Hong Kong Generative AI Research and Development Centre (HKGAI) and a HK$3 billion AI Subsidy Scheme. In addition, HK$1 billion has been earmarked for the establishment of the Hong Kong AI Research and Development Institute, which will spearhead both AI research and industrial applications under the oversight of the DPO. These efforts are complemented by talent development programs such as AI-focused university curricula, vocational training, and the Technology Talent Admission Scheme (TechTAS), which facilitates the recruitment of global AI professionals.

AI empowering engineers

AI is not replacing engineers—it’s empowering them. By automating routine tasks and providing real-time insights, AI allows engineers to focus on higher-level problem-solving and innovation. For instance, digital twins enable engineers to simulate production scenarios, test improvements, and predict outcomes without disrupting actual operations.

Moreover, machine learning algorithms help engineers identify inefficiencies, optimise resource allocation, and forecast equipment failures. Computer vision systems assist in quality assurance, while NLP tools streamline documentation and communication. These capabilities enhance decision-making, reduce human error, and foster a more agile and responsive manufacturing environment.

The benefits of AI in manufacturing

The benefits of AI in manufacturing are substantial. AI enhances efficiency by optimising workflows and reducing downtime. It improves product quality through real-time monitoring and predictive analytics. Cost savings are achieved via predictive maintenance and energy optimisation. AI also enables mass customisation, allowing manufacturers to adapt quickly to changing consumer demands. Importantly, AI contributes to sustainability by minimising waste and energy consumption.

AI offers a wide range of benefits. For example:

• Increased efficiency: Optimises workflows, reduces cycle times, and minimises human error.
• Improved quality: Real-time monitoring and predictive analytics ensure consistent product quality.
• Cost reduction: Predictive maintenance and resource optimisation lower operational costs.
• Customisation: Enables flexible production lines that can adapt to changing customer demands.
• Sustainability: Reduces energy consumption and material waste.

The challenges of AI in manufacturing

Despite its potential, AI adoption is not without challenges. Data silos and legacy systems hinder seamless integration. There is a shortage of skilled talent capable of bridging AI and industrial domains. Cybersecurity risks rise with increased connectivity, while ethical concerns—such as algorithmic bias and decision transparency—must be addressed.

Overcoming these challenges requires coordinated efforts among governments, industry leaders, and academia to build robust infrastructure, develop talent pipelines, and establish clear governance frameworks.

Key challenges:

• Data silos: Fragmented data systems hinder AI training and integration.
• Talent shortage: There is a growing need for engineers and data scientists with domain-specific knowledge.
• Integration complexity: Legacy systems often lack compatibility with modern AI platforms.
• Cybersecurity risks: Increased connectivity raises concerns about data breaches and system vulnerabilities.
• Ethical and regulatory Issues: Ensuring transparency, fairness, and compliance is critical.

Emerging trend of AI in manufacturing: Providing the platform solution

The manufacturing industry is undergoing a profound transformation driven by the integration of AI technologies. As manufacturers seek to enhance efficiency, reduce costs, and improve product quality, AI emerges as a pivotal force, reshaping traditional processes through innovative solutions. One of the most significant trends in this evolution is the rise of platform solutions that leverage AI capabilities, focusing on two critical aspects: fast, secure, and customisable platforms, and the deployment of local AI agents powered by large language models (LLMs).

Fast, secure, and customisable

In the competitive landscape of manufacturing, speed is paramount. AI-driven platforms are designed to streamline operations, enabling manufacturers to respond swiftly to market demands and optimise production workflows with unprecedented agility. By analysing vast amounts of data in real-time, AI algorithms provide actionable insights that support rapid decision-making. Through predictive maintenance frameworks and the automation of routine tasks, AI also reduces downtime and boosts overall operational efficiency.

Security is equally crucial. As AI becomes embedded in manufacturing systems, the protection of sensitive data is a top priority. Modern AI platforms incorporate advanced security measures, including encryption, access controls, and anomaly detection, to safeguard proprietary information and ensure the integrity of production processes. This combination of speed and security builds trust and allows manufacturers to accelerate digital transformation with confidence.

Customisation adds another critical dimension. Every manufacturing environment is unique, and AI solutions offer flexibility to tailor functionalities to specific needs. Customisable AI platforms enable manufacturers to configure algorithms, dashboards, and reporting tools to align with operational requirements and strategic goals. This adaptability ensures that AI solutions are not generic but are finely tuned to drive productivity and innovation in specific context.

Local AI agent - Large Language Model (LLM)

The deployment of local AI agents, particularly those powered by large language models (LLMs), represents a breakthrough in applying AI to manufacturing. With their capability to understand and generate human-like text, LLMs offer transform communication, knowledge management, and decision support in complex manufacturing industry settings.

Local AI agents, embedded within manufacturing facilities, provide real-time assistance to workers and managers. These agents leverage LLMs to interpret complex data, generate reports, and deliver data-driven recommendations based on historical and current information. For instance, a local AI agent can analyse production data to identify inefficiencies, suggest process improvements, or even predict equipment failures before they occur. By providing these insights, LLM-powered agents empower manufacturers to make informed decisions that optimise operations and reduce costs.

Moreover, local AI agents enhance collaboration and communication by translating technical jargon into accessible language, enabling cross-functional teams and workers at all levels to engage with and utilise complex data effectively. This democratisation of information fosters a culture of continuous improvement and innovation, as employees can leverage AI insights to drive process enhancements and product development.

The integration of LLMs also extends to customer interactions. Manufacturers can deploy local AI agents to manage inquiries, provide product information, and address concerns, delivering personalised, human-like responses, and strenghten customer loyalty.

Application cases of AI in manufacturing platform solution

3D Defect Detection for High-Mix-Low-Volume Products

In manufacturing high-mix-low-volume products such as footwear, AI-driven platform are revolutionising quality control. Traditional inspection systems are time-consuming and poorly suited to the diversity of product standards. For example, the time-intensive process of inspecting each shoe manually can take about half an hour using conventional methods. Moreover, commonly used visual AI techniques often lack sensitivity to 3D defects and complex product geometries, limiting their effectiveness.

Figure 6: 3D Defect Detection for Shoes (Courtesy of HKPC)

Figure 7: AI-enabled smart negative pressure palmier transportation system (Courtesy of CM Bakery Production Limited)

To address these challenges, an innovative approach combines 2D and 3D inspection methods with real-time AI analysis. This dual-method strategy significantly enhances the adaptability and efficiency of defect detection systems. By integrating advanced scanning technology with AI, manufacturers can now scan each shoe in just 5–6 seconds, dramatically reducing inspection time while boosts productivity and throughput.

This AI-enhanced platform rapidly adapts to new shoe models, leveraging its ability to detect defects in real-time. The system's agility in handling ever-changing product shapes and sizes ensures that manufacturers maintain high quality standards without compromising speed. By utilising both 2D and 3D data, the AI system achieves a comprehensive understanding of each product's unique contours and dimensions, enabling precise identification of defects that might otherwise go unnoticed.

This case exemplifies how AI in manufacturing platform solutions can effectively address the pain points of traditional inspection systems, offering a scalable, efficient, and adaptable alternative that meets the demands of modern production environments.

AI-enabled smart negative pressure transportation system

In traditional palmier packaging, manual alignment and placement, often leads to high defect rates, with fragile pastries breaking during handling, resulting in product waste and diminished customer satisfaction. To address this, a novel AI-enabled smart negative pressure transportation system has been developed to fully automate the pick-and-place process, eliminating manual intervention while improving yield and efficiency.

The system’s core is a high-precision negative pressure device, which uses finely calibrated nozzles to apply controlled suction. Sensors and actuators regulate the air pressure and maintain the ideal engagement height for each pastry, while closed-loop PID controllers ensure a consistent vacuum level. This reduces mechanical stress and effectively prevents cracking or damage to the delicate products.

AI technologies enhance both accuracy and adaptability. Computer vision delivers real-time spatial awareness and object recognition, feeding data to a machine learning model trained on extensive palmier image datasets. The model extracts key features such as edges, contours, and orientation to guide precise picking. Advanced feedforward control anticipates necessary adjustments based on the pastry’s initial position and orientation, while feedback control continuously corrects deviations during the process, ensuring that each palmier is securely grasped and accurately placed into its packaging compartment.

By integrating automated negative pressure handling with AI-driven precision, the system significantly reduces defect rates, improves packaging efficiency, and lowers costs. It also boosts yield by cutting waste from damaged products, while its consistent product quality enhances customer experience. This AI platform solution sets a new benchmark for fragile food packaging, demonstrating how robotics, AI, and advanced control systems can transform manual processes into reliable, high-performance operations.

Role of engineers in AI in manufacturing

In the engineering sector, there remains a clear gap in AI development skills. Building stronger links between AI experts and engineers is crucial to address this shortfall. Engineers play a pivotal role in integrating AI into manufacturing, serving as the bridge between advanced technologies and practical applications, as illustrated in Body of Knowledge (BoK2) for Engineers. Their expertise is crucial in three key areas: domain knowledge, human-machine collaboration, and safety and compliance.

Domain knowledge

Engineers bring deep domain knowledge to the table, which is essential for tailoring AI solutions to the specific requirements of different manufacturing contexts. Their understanding of production processes, machinery, and materials enables them to identify high-impact areas where AI can optimise operations, such as improving quality control, reducing waste, and enhancing supply chain efficiency. This expertise ensures that AI implementations are not only technically robust but also strategically aligned with the operational goals and challenges of the manufacturing environment.

Human-Machine Collaboration

Engineers play a vital role in fostering effective human-machine collaboration. They design systems that integrate seamlessly with human workflows, ensuring that AI tools augment rather than replace human capabilities. By developing intuitive interfaces, providing comprehensive training, and prioritising user-centred design, engineers enable workers to harness AI insights to improve decision-making and efficiency. This collaboration creates a more productive and engaged workforce, where humans and machines work in synergy to capitalise on their respective strengths.

Safety and Compliance

Safety and compliance are paramount in manufacturing, and engineers play a central role in ensuring that AI systems meet these critical standards. They carry out rigorous risk assessments to identify potential hazards associated with AI deployment and implement safeguards to protect workers, equipment, and data. Additionally, engineers ensure that AI solutions comply with industry regulations, data privacy laws, and safety standards, maintaining both the operational integrity and organisational reputation. Through meticulous planning, thorough testing, and ongoing oversight, engineers ensure that AI strengthens safety and compliance, fostering a secure and reliable manufacturing environment for all stakeholders.

Conclusion: Empowering new quality productive forces with AI in manufacturing platform

The journey ahead requires vision, collaboration, and investment. But the destination—a smarter, more agile, and more inclusive manufacturing ecosystem—is well within reach.

AI is not just transforming manufacturing—it is redefining the very nature of production. In Hong Kong, the integration of AI into manufacturing platforms is creating a new paradigm of intelligent, efficient, and sustainable industry. These New Quality Productive Forces (NQPFs) are essential for building a resilient economy that can adapt to global challenges and seize emerging opportunities.

At the heart of this transformation are engineers, whose domain expertise, problem-solving skills, and commitment to safety and compliance are more crucial than ever. AI is not replacing engineers—it is empowering them. By automating routine tasks and delivering real-time insights, AI enables engineers to focus on higher-level innovation and strategic decision-making.

By embracing AI, Hong Kong can:

Empowering New Industrialisation : The Future is Engineered!

Figure 8: AI in Body of Knowledge (AI -BoK2) for Engineers & Technologists to engage in New Industrialisation

Technologies such as digital twins, machine learning, computer vision, and natural language processing allow engineers to simulate, optimise, and streamline operations—enhancing decision-making and reducing errors. Yet, this shift also brings challenges: engineers must adapt to new tools, ensure ethical AI use, and uphold safety and compliance in increasingly complex systems.

The journey ahead requires vision, collaboration, and investment. But the destination—a smarter, more agile, and more inclusive manufacturing ecosystem—is well within reach. With AI as a cornerstone, and engineers as its stewards, Hong Kong is poised to lead the next wave of industrial excellence in the digital age.

About the author

Ir Raymond Shan is the Chairman of the HKIE MI Division. He is also the General Manager, New Industrialisation Division of Hong Kong Productivity Council.

Reference

World Economic Forum (WEF). (2025). Frontier Technologies in Industrial Operations: The Rise of Artificial Intelligence Agents White Paper January 2025. Retrieved from WEF_Frontier_Technologies_in_Industrial_ Operations_2025.pdf