[The Road to M.AX②] AI Becomes the 'Eyes' and 'Brain' of the Factory

Beyond 'Automation' to 'Decision-Making'... How AI Is Transforming the Factory Landscape

Minister of Trade, Industry and Energy Kim Junggwan is listening to an explanation about the intelligent autonomous maintenance integrated system from an official at Hanla IMS, a shipbuilding equipment company located in Hwajeon Industrial Complex, Gangseo-gu, Busan, on May 27. Ministry of Trade, Industry and Energy.

Just a few years ago, artificial intelligence (AI) in the manufacturing industry was limited to automating repetitive tasks. Now, however, AI assists with research and development (R&D), detects anomalies before equipment failures occur, and identifies defects in place of skilled workers’ eyes. It has evolved into a stage where it supports decision-making by learning from decades of accumulated human experience. Manufacturing AI Transformation (M.AX) is no longer a concept of the future, but a change already underway in manufacturing sites.

According to cases from the Industrial AI Solution Demonstration Project supported by the Korea Institute for Advancement of Technology (KIAT), reported on July 8, the direction of AI was similar across different industries, including semiconductors, automobiles, shipbuilding, steel, and chemicals. AI was not limited to automating a single process within a factory, but was expanding its reach across the entire manufacturing process, from R&D to production and quality control.

Faster R&D, More Stable Production Lines

The first place to change was not the production line, but the research lab. Jusung Engineering, a semiconductor equipment company, has implemented a generative AI system based on its internal closed network to support R&D. Researchers now use AI to instantly search for and utilize the vast amount of technical documents and design data needed in the design process, instead of searching for them one by one. This has reduced the time spent on repetitive data searches, allowing researchers to focus more on research and design. The company is taking this a step further by upgrading its system so that AI will be able to predict experimental results.

A KIAT official explained, "There are more cases where manufacturing AI achieves results in the R&D stage before the production line, and AI is evolving to utilize companies’ technological assets."

On the production line, AI now monitors equipment. Hwaseen, an automotive parts manufacturer, has introduced an AI-based predictive maintenance system into its electric vehicle battery pack case production process. The AI analyzes real-time data collected from sensors attached to the equipment and detects early signs of anomalies. Whereas previously, workers would find the cause and repair the equipment after it had already stopped, now the system notifies them of maintenance timing before breakdowns occur.

This has led to measurable results. The AI can predict equipment anomalies up to 72 hours in advance, and workers are now able to replace tools in a timely manner, virtually eliminating unexpected stoppages. Over the next five years, Hwaseen expects to save approximately 2.8 billion won in costs and secure a foundation for new orders worth around 250 billion won.

AI Begins to Learn the Human Eye

Quality inspection is one of the manufacturing processes that requires the highest level of skill. It takes years of experience to spot even a minor flaw, and inspection results can vary depending on the worker.

Donghwa Entec, a shipbuilding equipment company, has established a system in which AI analyzes shape data generated during the heat exchanger welding process in real-time to predict quality. Moving away from the previous method of checking for defects after production is complete, the company now detects quality issues during the manufacturing process itself. Since implementing AI, inspection time has decreased by about 30%, and the need for rework has been reduced by about 20%.

Yulchon Chemical, a chemical company, has also introduced an AI vision inspection system. Previously, workers had to manually check thousands of clips during the film production process, but now AI analyzes them instead. Inspection time has been cut to under 30 minutes, and the reading accuracy has improved to over 95%. The company estimates that this will reduce annual losses by around 1 billion won.

The role of AI is not limited to production processes. AI is now being introduced to raw material management, the starting point of manufacturing. YK Steel, a steel company, has adopted an AI-based steel scrap vision solution, transforming its raw material inspection system. Scrap inspection had long relied on workers’ experience, but now AI automatically determines the type and grade of incoming materials, shifting to a data-based inspection system.

After implementing AI, the consistency rate between AI judgment and visual inspection has increased to over 90%, and five workers can now manage ten yards collectively. Disputes over grading with suppliers have also decreased. This not only improved work efficiency but also strengthened supply chain reliability.

It Is Now Time for Manufacturing AI to Spread

Although the industries are different, these cases share a common thread: AI has begun to take on the judgment tasks that previously consumed the most time by humans, rather than simply replacing people. It is now helping researchers with design, predicting equipment failures, assessing quality in advance, and analyzing raw materials. Rather than being just a process automation tool, AI is evolving into the “eyes” and “brain” of the factory.

However, many small and medium-sized enterprises are struggling to implement AI due to issues such as insufficient funding, lack of skilled professionals, and data quality problems. This is why the government is promoting the M.AX Alliance and MINI Alliance—to spread AI innovation throughout industrial complexes, partner companies, and the entire regional manufacturing ecosystem.

KIAT President Jeon Yoonjong said, "AI is not just for a few high-tech companies. The key is to improve workflow to accumulate data, and, through this, enhance efficiency and productivity. We will actively promote AI solutions tailored to manufacturing companies and vigorously support the creation of an ecosystem for sharing and utilizing manufacturing-related information and resources to solve common industry challenges."

This content was produced with the assistance of AI translation services.