Driven by production to support education: Is it the "savior" of the AI + manufacturing industry?
In 2025, it is regarded as the first year of AI agents.
As the imagination space for agents has expanded from "assistants" to actual positions and "digital employees," higher expectations have been placed on the significance of AI for the digital transformation of the manufacturing industry.
Especially in the manufacturing industry, behind the gap between the proliferation of digital systems and the failure of most enterprises to achieve digital transformation, a structural contradiction is reflected: enterprises have deployed new digital applications, but no one knows how to use them.
According to statistics from e-workers Research, among the current employees in the manufacturing industry, only 2% of enterprises have talents with AI professional skills and knowledge. However, AI must grow in the soil of industrial practice, rather than floating in the cloud of technical architecture. In the case of a defective AI + talent cultivation system, the implicit investment in reshaping personnel capabilities during the digital transformation of the manufacturing industry is even higher than the technical cost.
"In any era, the real competition is the competition for talents," Wei Xiaogang of Midea Group's Meiyun Zhishu told 36Kr.
A few days ago in Shenyang, Meiyun Zhishu initiated the Manufacturing Digital Transformation ME20+ Forum (Manufacturing Education 20+ Forum, abbreviated as ME20+ Forum) for the first time and released the ME Agent Co - creation Plan (Manufacturing Education Agent). This also reveals the "new rule" for the digital transformation of the manufacturing industry - talents as a service.
A Shortage of 5 Million AI Talents May Slow Down the Manufacturing Industry
What defeats people is not AI, but those who can use AI.
In the wave of digital transformation in the manufacturing industry, a consensus is emerging: what really hinders the progress of the industry is not the technology itself, but the people who master the technology. According to a McKinsey survey, more than 70% of digital transformation projects fail to meet the expected goals, mainly due to three reasons. First, enterprises on average have too many application systems, but integration and data circulation are blocked. Second, traditional automation processes cannot adapt to business changes, and the maintenance cost is high. Third, there is a significant shortage of digital talents.
The white paper released by the National Industrial Information Security Development Research Center shows that the shortage of artificial intelligence talents in China reached 300,000 in 2020, and the shortage in the field of intelligent manufacturing will climb to 5 million this year.
It can be seen that the transformation of the manufacturing industry is no longer the traditional "replacing humans with machines," but has shifted to "intelligent drive" completed by humans, so it deeply relies on the upgrading of personnel quality.
However, the compound talents in the manufacturing industry are not simply a stacking of technologies. They need to have a good understanding of information technologies such as industrial engineering, operational technology, and AI. Such talents need to understand the process pain points in the production process and be able to transform AI algorithms and industrial big data into practical cost - reduction and efficiency - improvement solutions in the workshop. At present, they are rare.
Although medium - and large - sized enterprises are developing their own digital talent cultivation systems, the internal self - sufficiency model has obvious limitations. First, it takes a long time. It takes at least two or three years from learning to business integration. Second, the cultivation cost is high. Third, there is a risk of talent loss. Digital talents in the manufacturing industry may flow to high - premium industries such as the Internet.
The "dilemma" of these superimposed factors is further magnified in the industrial chain collaboration. If the digital foundation of upstream suppliers is weak, it will be difficult for downstream enterprises to build a full - chain intelligent model. The shortage of talents is like a domino effect, slowing down the intelligent process of the entire manufacturing industry.
The deeper contradiction lies in the cultivation mechanism. Under the traditional education system, engineering students lack data thinking training, and AI + talents have a vague understanding of production lines and engineering practices. There is a "gap" between university education and enterprise needs. Today, the talent cultivation echelon and system for "AI + manufacturing" are almost uncharted territory.
To solve this dilemma, the key lies in building an ecosystem of school - enterprise collaboration, that is, the integration of industry and education.
Based on this, Meiyun Zhishu, together with more than 10 benchmark manufacturing enterprises in digital transformation and more than 10 universities, initiated the Manufacturing Digital Transformation ME 20+ Forum and the ME Agent Co - creation Plan for the first time. Centered on Midea Group's "lighthouse practice" in AI transformation, it enables enterprises and the education industry to "co - cultivate" digital and intelligent talents, achieving the effect of direct employment and scientific research collaboration. Under the theme of the integration of industry and education, enterprises build a "talent reservoir," while universities are the "source of living water."
At the same time, the advantage of the integration of industry and education is that it can transform the real enterprise scenarios into teaching courses, allowing students to complete the full - process training from algorithm development to process optimization in a twin factory. By reconstructing the talent evaluation system, it realizes the seamless connection of "starting work right after graduation," thus significantly shortening the talent cultivation cycle.
The Integration of Industry and Education Bridges the "Last Mile" of Talent Supply and Demand
The structural contradiction of talents is not unique to the digital age, but it has evolved into an obvious pain point after the popularization of AIGC. Especially after all industries have deployed underlying capabilities such as Deepseek, the productivity value provided by the same set of large - model applications varies significantly. The essence is still the gap in people's abilities.
In the era of tool equality, the "last mile" of value release completely depends on the user's business understanding and scenario reconstruction ability. Therefore, the ability gap between people is even more fatal.
It is precisely based on the problem of misalignment in AI + talent cultivation that Meiyun Zhishu initiated the ME20+ Forum, trying to close the gap from a single point. This path is not so much an exploration of a new solution as an industrial - level replication of Midea's digital talent system, which has been accumulated for many years, in the school - enterprise co - creation project, jointly building the "minimum viable unit" of the talent model.
The enterprises and universities covered by ME20+ are far more than the 20 involved this time. In the future, the "ecological circle of friends" will be expanded, and a co - creation mechanism with high precision, high coupling, focusing first and then spreading will be built. It will continuously "package AI capabilities" from scenario - based cultivation to practical verification, so as to standardize and integrate the knowledge systems originally scattered in universities, leading enterprises in the chain, and upstream and downstream enterprises, and then form a compound efficiency of talents flowing into the industrial chain.
In the future, after restructuring the industrial genes, the ME20+ Forum will extract the first batch of benchmark cases of the integration of industry and education, and popularize the AI + talent cultivation path of "manufacturing + education" on a large scale. In fact, the first benchmark case is in Shenyang Institute of Technology, the starting place of the ME20+ Forum.
The "digital and intelligent talent cultivation" system of Shenyang Institute of Technology is an exploration of the integration of industry and education aiming to "break the situation" for the cultivation of compound talents.
In the past few years, Shenyang Institute of Technology has established 18 modern industrial colleges and 24 experimental centers. The industrial colleges include those of Huawei, KUKA, Midea, etc. The curriculum system is jointly designed and taught by college teachers and enterprise engineers, tying up and meeting the real needs of the industry by breaking the teaching staff structure.
What is more breakthrough is the Midea Digital Lighthouse College just completed at Shenyang Institute of Technology. As a key node of Meiyun Zhishu's ME20+ Co - creation Plan, the Lighthouse College connects the modern industrial colleges and experimental centers of Shenyang Institute of Technology through a digital platform, carrying out integrated intelligent management, control, and evaluation in teaching and practice, so that the digital talent cultivation system can form a "task - driven practical teaching" based on the practical scenarios of the business.
Shenyang Institute of Technology has created a "showroom" for Midea's Lighthouse College. Its value lies not only in improving the efficiency of industrial education but also in concretizing the traditional and vague concept of compound talents with the help of AI.
For example, by tracking the students' practice data in projects such as energy consumption optimization and equipment predictive maintenance, a digital talent portrait containing multiple indicators is formed. The "pressure test" at the industrial level is used to bridge the "last mile" of talent supply from universities to the industrial end.
It can be said that Shenyang Institute of Technology is not just representing the education circle in embracing the industry, but implanting education into the process of industrial upgrading in an open and self - reforming manner, so that the "way to break the situation" for digital talents in China's manufacturing industry returns to education itself.
It is not difficult to understand why leading manufacturing enterprises are willing to co - create with Shenyang Institute of Technology. First of all, private universities are seizing the time difference in talent acquisition for the industrial upgrading of the entire manufacturing ecosystem with the flexibility of their curriculum and management mechanisms. Compared with public universities, which are restricted by the discipline evaluation system and invest more in scientific research and laboratories for AI + applications, the talent cultivation model of private universities can respond seamlessly to enterprise needs.
Midea Zhishu Lighthouse College disassembles enterprise - level technical solutions such as industrial simulation, digital twin, and AIGC into modular capabilities, allowing students to apply them in any engineering scenario design. Therefore, although the technical foundation of this education model comes from the tool applications, talent cultivation methods, and industrial cognition of leading enterprises themselves, its value spill - over effect has far exceeded the needs of a single enterprise. In essence, it sends talents into the industrial chain. While helping young people master emerging technologies earlier and opening up "future employment channels," it also installs an agile and sustainable "talent engine" for the intelligent development of the manufacturing industry.
In the "deep water area" of the integration of industry and education, AI itself is also an indispensable key. For example, the ME Agent Co - creation Plan released by Meiyun Zhishu, that is, Manufacturing Education Agent, actually uses the technical platform to make AI agents a talent service tool. Students can learn, ask questions, conduct ability certification and evaluation on it, and then match employment positions. Different from the traditional AI - assisted teaching system or the current assistant - type agents, the "learning, practice, evaluation" system of ME Agent is directly linked to employment ability, making intelligent courses, intelligent talent evaluation, and intelligent employment a "teaching symbiotic entity," interacting with students like professional tutors.
In this way, AI is not only the goal of students' self - iteration but also the bridge for the integration of industry and education.
In the Age of Agents, the Dynamic Is the Cutting - Edge
In the era of AI applications, a key contradiction in the integration of industry and education is that the rapid iteration of market demand requires a dynamic approach to talent cultivation, while the industrial advantages, practice scenarios, and cultivation directions in universities are relatively fixed.
Once there is a risk of "dynamic mismatch" between university teaching and enterprise practical applications, a group of students may "be unemployed right after graduation."
For example, if the enterprises need engineers who can fine - tune large models four years later, but students are still learning traditional machine learning frameworks, even "hard - core" AI talents with solid theoretical knowledge may find that their practical work experience is not enough to "be favored by enterprises." The breakthrough of the ME Agent Co - creation Plan is reflected here: by perceiving enterprise needs through the industry's dynamic knowledge base, converting them into teaching tasks, and finally precipitating them into knowledge, the students' training data can feed back into the enterprise's process optimization. This two - way penetration mechanism also enables digital talents to indirectly participate in enterprise innovation while mastering technologies.
On the other hand, the model itself can dynamically adjust the weight of the talent portrait according to the position requirements in the industry, and universities can optimize the curriculum modules according to the matching degree to improve the employment matching rate. At present, AI agents do not have the ability to "create courses" independently. Therefore, Meiyun Zhishu needs to continuously invest in converting "enterprise courses" into "teaching courses," "feeding" the large model with cutting - edge data, waiting for the next singularity moment.
Moreover, whether it is the ME20+ Forum or the process of universities such as Shenyang Institute of Technology choosing industrial colleges, it is inseparable from the tracking and implementation of national and regional talents. The policy drive of the national digital economy plan is the driving force for the industrial upgrading of the manufacturing industry, and it also continuously gives birth to a number of emerging positions, such as agent creators and marketing engineers.
For a student with a long - term career development plan, if they want to transfer their abilities from general positions to scenario - based collaboration, the most important thing is to infer the skills and resumes they need to supplement from employment prediction. For example, a marketing student has many fields to choose from. Only by predicting the possible marketing needs of enterprises four years later can they evaluate their interests and "ability map," and all this can be evaluated through ME Agent.
In fact, the integration of industry and education is not limited to the model of Shenyang Institute of Technology. In the past, enterprises carried out targeted training in universities, aiming to enable students to "start work right after graduation." The "digital and intelligent talent cultivation" model of Shenyang Institute of Technology not only combines the capabilities of multiple enterprises and realizes individualized teaching through AI Agent but also helps students get rid of the "vertical ability trap," so that even if graduates cannot enter leading enterprises, they can still find employment in their ecological chains.
Therefore, to prevent the integration of industry and education from becoming rigid, it is also very important to compress the digital methodology of the manufacturing industry into relatively general - purpose ability components, allowing students to improve their own ability stacks "like building Lego," activate their innovation awareness, and find employment in very different industries. Only in this way can the single - to - single "targeted blood transfusion" be transformed into "ecological blood creation."
In fact, the essential difference between manufacturing enterprises and factories lies in whether they can penetrate business operation thinking into all value stages, including R & D, production, supply chain, and service. In the AI era, the equipment competition in many industries has completely shifted to a competition for talent density. The active pursuit awareness of human - machine collaboration largely determines a person's value and also affects future competition.
