Intelligent Manufacturing Insights Column | Liu Zhifeng: Artificial Intelligence Empowering Intelligent Manufacturing — Current Status, Trends, and Challenges

Artificial intelligence (AI), as the core engine of the new round of technological revolution and industrial transformation, is reshaping the global innovation landscape through a "leading goose effect." General Secretary Xi Jinping has insightfully pointed out that development should "take intelligent manufacturing as the main focus to drive industrial technological transformation and optimization and upgrading," and he has explicitly emphasized "promoting the deep integration of AI scientific and technological innovation with industrial innovation." In June 2025, the Ministry of Industry and Information Technology, while reviewing the "Work Priorities for Integration of Informatization and Industrialization 2025," explicitly proposed implementing an "AI + Manufacturing" initiative to accelerate intelligent upgrades in key industries and create an upgraded version of intelligent manufacturing.

1. AI empowering manufacturing: notable progress in intelligent manufacturing

As a key driving force for intelligent manufacturing, artificial intelligence has made significant breakthroughs in frontier areas such as technology integration, reconstruction of production models, and innovation of industrial ecosystems. Through deep integration with advanced technologies like industrial robots, machine vision, and digital twins, AI is helping manufacturing break free from the constraints of traditional production models, accelerating the shift from automation and digitalization toward intelligence, and from single-machine intelligence toward full-chain intelligent collaboration. The specific progress is reflected in the following three aspects:

Technology integration: breakthrough applications of intelligent equipment and sensing systems

The deep integration of AI technologies with manufacturing equipment and sensing systems has overcome the inherent limitations of traditional production devices that only “passively execute,” establishing a closed-loop intelligent system of “perception–decision–execution.”

In the field of intelligent equipment, taking AI-driven welding robots as an example, multi-dimensional data such as temperature, current, and weld seam morphology are collected in real time during the welding process, and welding parameters are dynamically adjusted based on this data, effectively reducing welding defect rates. Flexible collaborative robots, leveraging AI visual positioning technology, can accurately identify the position and pose of different component models to complete assembly operations, highly adapting to multi-variety, small-batch production demands.

Machine vision systems integrated with deep learning algorithms are no longer limited to simple dimensional measurement and appearance defect detection; they can accurately recognize and classify product defects in complex scenarios, quickly and precisely detecting latent defects such as scratches and contaminants, ensuring production efficiency and product quality.

The organic combination of generative AI and intelligent equipment has realized “reverse design” of production processes. By inputting product performance requirements, AI can automatically generate optimal equipment processing paths and parameter configurations, helping enterprises shorten product development cycles.

Reconstruction of production models: from automation to autonomy

The deep penetration and widespread application of artificial intelligence technology is driving the manufacturing production model to shift from "mechanical automation" toward "autonomous intelligent production", achieving "self-perception, self-decision-making, and self-optimization" of the production process.

In production scheduling within intelligent factories, AI systems collect real-time data on equipment operating status, raw material inventory, order demands, and other factors from the production line, and use multi-objective optimization algorithms to dynamically adjust production plans and equipment rosters. Even if equipment fails, AI can automatically reassign production tasks for the equipment and simultaneously generate fault repair plans to avoid production interruptions; when raw material supply is delayed, AI can predict arrival times based on historical data, proactively adjust production sequences, and prioritize production of products with sufficient inventory to ensure order fulfillment rates are not affected.

In quality control, AI systems transform the production mode from "post-inspection" to "prevention", building quality prediction models by analyzing in real time the relationships between key process parameters and product quality data. When parameters show abnormal trends, warnings can be issued in advance and production parameters automatically adjusted to prevent the generation of nonconforming products.

AI systems also better enable "flexible customization" of production. By interfacing AI with customer demand management systems, customer personalization requirements can be quickly converted into production instructions, enabling "single-piece ordering, mass production" to fully meet consumers' customized demands.

Industrial Ecosystem Innovation: Full-Chain Intelligent Upgrading

AI systems have enabled intelligent upgrades across the entire chain—R&D, production, supply chain, sales, and service—driving manufacturing from "single-machine intelligence" to "full-chain collaborative intelligence" and creating a brand-new industrial ecosystem.

In the R&D phase, the integration of AI and digital twin technologies has created a closed-loop system of "virtual R&D–real validation." R&D personnel can build virtual models of products on a digital twin platform and use AI to simulate product performance under different operating conditions, rapidly optimize product designs, reduce the number of physical prototypes, significantly shorten R&D cycles, and lower R&D costs.

In the supply chain phase, AI-driven supply chain collaboration platforms enable real-time information sharing and intelligent interaction among upstream and downstream enterprises. Based on market demand forecasting, they guide upstream suppliers to adjust raw material production plans to avoid inventory buildup, and use AI logistics scheduling algorithms to optimize transportation routes and warehouse layouts, reducing supply chain logistics costs.

In sales and service, AI technologies can be used to create precise customer profiles, analyze customers' consumption habits and preference patterns, and enable targeted marketing and personalized recommendations; in after-sales service, AI remote diagnosis systems can use data collected from equipment sensors to monitor equipment operating status in real time, predict equipment failures in advance, and remotely guide customers through repairs or dispatch nearby service personnel for on-site service, greatly improving after-sales service efficiency.

In the recycling and processing phase, machine learning can predict community recycling volumes and optimize route-based dynamic scheduling of collection vehicles, RFID tags can track the flow of bulk discarded items; computer vision systems can identify waste materials in real time, spectral analysis techniques can detect toxic and hazardous substances; image recognition technology can assess the wear level of second-hand goods, blockchain technology can trace and verify high-value components, and market big data analysis can generate real-time recycling price guidance.

2. Deep integration of artificial intelligence and manufacturing: a promising future for intelligent manufacturing

Multi-technology integration

The development of intelligent manufacturing cannot rely solely on advances in a single AI technology; it requires the deep integration of multiple cutting-edge technologies such as AI, sensors, IoT, big data, cloud computing, and the advancement of manufacturing technologies themselves. Advances in sensor technology have made the states of equipment, materials, and factories observable and measurable; IoT has enabled interconnection among production equipment, products, and raw materials, allowing realtime data collection and transmission so the entire production system becomes “everything connected”; big data provides AI with massive and rich data resources, and through deep mining and analysis of these data, AI models can accurately detect potential problems and optimization opportunities in the production process; cloud computing provides powerful computing support for data storage and processing, ensuring that largescale data computations can be performed efficiently. Finally, AI can use this foundation to optimize production processes and predict equipment failures, achieving seamless integration and efficient operation of production lines, and comprehensively enhancing the intelligence and efficiency of manufacturing.

Global Collaboration and the Formulation of Regulatory Standards

Intelligent manufacturing involves the reshaping and upgrading of national manufacturing industry chains and requires extensive global cooperation and the formulation of scientifically sound standards. In terms of technical exchange, researchers and companies from different countries can share cutting-edge technologies and innovations in AI applications for manufacturing, accelerating the global dissemination and iteration of these technologies. The formulation of standards is even more critical: unified technical standards and norms, such as communication protocols and data formats, can ensure good compatibility and interoperability among equipment and software systems produced by different countries and manufacturers, reducing integration problems caused by standard differences and comprehensively promoting the healthy, orderly, and high-quality development of intelligent manufacturing.

Towards Carbon Neutrality and Sustainable Development

With growing global attention to climate change and the ecological environment, and increasing public environmental awareness, green manufacturing has become a development direction that intelligent manufacturing must take seriously. At the energy management level, AI can monitor the energy consumption of each piece of equipment in real time and flexibly adjust equipment operating power and start-stop times based on production tasks, achieving precise allocation and efficient use of energy. AI can encode carbon emissions and energy consumption as optimization objectives into process “genes,” enabling closed-loop “carbon budget” management from design and scheduling to recycling; an edge–cloud collaborative energy–carbon operating system can turn factories into flexibly tradable “negative-carbon assets,” driving interoperability of carbon footprint data across the supply chain and making “zero-carbon orders” a new passport for exports to Europe and the U.S. Ultimately, this will help the manufacturing sector as a whole reach a carbon-peak plateau before 2030 and lay the industrial foundation for carbon neutrality by 2060.

3. Challenges Faced by AI Empowering Intelligent Manufacturing

Increasing difficulty of technical integration

Intelligent manufacturing systems are a large and complex ecosystem that embody multidisciplinary intersections and integrate various sophisticated technologies—such as mechanics, materials, automation, and computer science—as well as data interfaces, data formats, data communication, system compatibility, and security. In particular, some vendors use encrypted communication protocols and data formats to protect their technical advantages, making it difficult for AI systems to obtain correct information and constraining the integrated application of AI technologies.

Data Security and Privacy Issues

Data is a core asset, encompassing business secrets, production process data, customer information, and other critical content, making data security and privacy protection essential. As data flows and is shared across different stages and entities, the boundaries of data privacy become blurred. Ensuring that data is not misused or that user privacy is not leaked during use requires robust data security mechanisms and privacy protection technologies.

Incomplete technical standards and regulations

Currently, the application and development of intelligent manufacturing are progressing rapidly, but the formulation of related standards and regulations is relatively lagging. In terms of technical standards, the lack of unified norms means that AI applications, intelligent devices, and so on developed by different companies differ significantly in functionality, performance, interfaces, and other aspects, affecting the healthy development of the industry ecosystem; there is also a regulatory gap—for example, the issue of liability allocation when AI-driven decisions are involved. When an AI system makes errors in production decisions that lead to product quality incidents or production accidents, it is difficult to determine whether the algorithm developer, the data provider, or the enterprise itself should bear responsibility. This uncertainty increases companies’ reservations about applying AI technologies and restricts the promotion and scope of application of intelligent manufacturing technologies.

There are barriers to employees’ acceptance of AI

On one hand, company employees have doubts about the safety and reliability of AI technologies, worrying that the introduction of AI systems will replace their own jobs; on the other hand, employees question the stability of AI in production, fearing that equipment failures could trigger production accidents. This distrust of the technology reduces employees’ willingness to cooperate with AI in their daily work and may affect the adoption and deep application of AI technologies within the company.

4. Summary

AI-enabled intelligent manufacturing is not only a technical upgrade but a reshaping of the entire manufacturing ecosystem. AI technologies will accelerate the industry's development toward greater intelligence, efficiency, and sustainability. To respond to this trend, companies need to develop long-term strategies for intelligent transformation and digitalization, and strengthen cross-domain technology integration and talent cultivation; governments should accelerate the improvement of relevant laws and regulations to provide a favorable institutional environment for technological innovation and industrial application; industry organizations need to build public platforms for standards, testing, and certification to promote collaborative innovation across the upstream and downstream of the industry chain; universities, as the source of knowledge innovation and supply of high-end talent, should proactively align with industry needs, establish interdisciplinary "AI + manufacturing" programs, and continuously produce interdisciplinary talent in the AI + intelligent manufacturing field.

Only through coordinated efforts by all parties can we fully unleash the potential of artificial intelligence in intelligent manufacturing, seize the opportunities presented by the new wave of technological revolution, and drive the manufacturing industry toward high-quality, sustainable development.