[Report] 15-Year Veteran Field Mechanic Becomes AI Talent... Korean Air's Experiment [AI Era, Jobs Are Changing]

Launched in August 2023, the Only Team in Korea’s Aviation Industry
Proactive Maintenance Approach Enhances Both Safety and Cost Efficiency
Over 2,500 Sensors per Aircraft, Generating 62GB of Data Daily
Detecting Anomalies in Advance Using Big Da

Editor's NoteThe use of artificial intelligence (AI) in industrial settings, including manufacturing, is expected to become the biggest topic in the industry this year. AI, which has already been introduced to manufacturing sites, has proven to play a pivotal role in boosting productivity. This year, its scope will expand significantly. However, concerns about job losses lie beneath the surface of AI-driven production innovation. As the so-called 'Dark Factory' concept-where production occurs without human intervention-becomes more of a reality, the prospect of full automation is drawing closer. Yet, The Asia Business Daily has found that changes in jobs at manufacturing sites where AI is applied are complex. Machines are taking over repetitive and hazardous tasks, while humans are increasingly focused on decision-making, management, and responsibility, resulting in a clear division of roles. Rather than replacing people, AI is beginning to redefine jobs. The AI era brings not only competition in productivity and technology, but also the challenge of job transition. The Asia Business Daily visited industrial sites in the new year to observe firsthand how AI is transforming the job landscape.

On the 9th of last month, at the Korean Air maintenance hangar in Jung-gu, Incheon, the Korean Air predictive maintenance team and a technician are inspecting the temperature sensor installed at the engine intake of an A330 aircraft. Photo by Kang Jinhyung

"The mechanics often ask the predictive maintenance team if we have some kind of 'supernatural power.' They wonder how we can identify defects days in advance without even seeing the aircraft in person."

On December 9, at the Korean Air headquarters in Gangseo-gu, Seoul, an alert appeared on the computer monitor of Park Changhoon, Deputy Manager of the Predictive Maintenance Team. The alert indicated that the cabin pressure of a passenger aircraft en route to Gimpo was dropping. Sensors installed on the aircraft detected that the rubber connectors between the ducts (the hoses through which air flows in the cabin) had become loose. This signaled that, if left unaddressed, the duct connectors could tear within a week, potentially preventing sufficient air from being supplied to the cabin during flight. Park analyzed related data, including past maintenance records, and sent a work request to the on-site maintenance team, recommending that the duct connectors be replaced.

He said, "We were able to catch a risk that could have led to a return flight or cancellation early on, allowing us to perform maintenance before the next flight," adding, "Recently, similar defects have occurred three to four times due to factors such as component aging, so we were able to quickly identify the cause of the pressure drop."

On the 9th of last month, at the Korean Air maintenance hangar in Jung-gu, Incheon, the Korean Air predictive maintenance team and a mechanic are inspecting the temperature sensor installed at the engine intake of an A330 aircraft. Photo by Kang Jin-hyung

Predictive maintenance refers to a proactive maintenance approach that uses big data and artificial intelligence (AI) to detect and address aircraft issues in advance. Korean Air equips each aircraft with about 2,500 sensors, generating an average of 62 gigabytes (GB) of data per day-the equivalent of 63,000 e-books. The company analyzes this data with an in-house AI-based defect prediction model, and when signs of abnormality are detected, it assesses whether maintenance is needed and then issues work orders accordingly.

The reason predictive maintenance is more efficient than traditional maintenance methods is that it can avoid the time and costs associated with aircraft defects. If a defect is discovered during pre-flight maintenance, it usually results in delays or cancellations. Additional costs are incurred for accommodations and alternative flights during maintenance. However, predictive maintenance allows defects to be anticipated about a week in advance, securing the necessary maintenance time ahead of schedule and reducing costs.

On the 9th, at the Korean Air Maintenance Hangar in Jung-gu, Incheon, Korean Air predictive maintenance engineers and mechanics are inspecting the temperature sensor installed on the engine inlet of the A330 aircraft. Photo by Kang Jinhyung

In response, Korean Air established a dedicated Predictive Maintenance Team in 2023. This team, the only one of its kind in the Korean aviation industry, identified over 100 defects in advance this year and achieved cost savings amounting to several billion won.

For example, the Predictive Maintenance Team once detected abnormal signs in the landing gear using visual sensors. The aircraft's wheels, which should fold up to a 90-degree angle after takeoff, were flying in a bent position. Since this part requires lifting the entire fuselage for inspection, it is difficult to detect such defects early without predictive maintenance technology. If the wheel angle were to widen further in subsequent flights, the wheels might fail to deploy during landing, making a return flight unavoidable. The affected aircraft was a large model with 90 business class seats and a total capacity of 300 passengers, so a return flight would have incurred significant costs.

Oh Jonghoon, Head of the Predictive Maintenance Team at Korean Air, is being interviewed at the Korean Air Maintenance Hangar in Jung-gu, Incheon on the 9th of last month. Photo by Kang Jinhyung

The significance of the Predictive Maintenance Team lies in the fact that its members were not hired externally, but rather transitioned from field roles such as mechanics and engineers, and retrained to develop AI capabilities. Oh Jonghoon, Head of the Predictive Maintenance Team, said, "Our top priority is always someone who knows aircraft well," adding, "Because a deep understanding of aviation systems is essential, we prefer to recruit employees interested in AI from our internal pool rather than hiring AI experts from outside."

In particular, mechanics who have accumulated hands-on experience and know-how from years of working directly with aircraft are highly valued. Deputy Manager Park also worked as a Korean Air mechanic for 15 years. When the airline began suspending operations during the COVID-19 pandemic, he started learning programming to keep up with the 'coding boom.' His skills quickly improved to the point where he could easily build work-related websites and programs, and he joined the Predictive Maintenance Team when it was formed in August 2023. Park said, "While working as a mechanic, I often wondered how I could reduce repetitive tasks and work more efficiently. After encountering data automation through programming, I became interested in coding," adding, "When I heard that a predictive maintenance team would be formed after the pandemic, I wanted to be part of it."

On the 9th of last month, at the Korean Air maintenance hangar in Jung-gu, Incheon, the Korean Air predictive maintenance team and a technician are inspecting the temperature sensor installed on the engine inlet of an A330 aircraft. Photo by Kang Jinhyung

Assistant Manager Kim Jaemin, formerly an engineer at Korean Air, was responsible for analyzing the causes of technical problems during aircraft operations and developing preventive measures. Kim said, "Predictive maintenance appealed to me because it focuses on anticipating and resolving issues before defects occur, rather than after the fact," adding, "I studied coding on my own after work, striving to create synergy with the practical know-how I gained in the field."

Korean Air has also made efforts to support the coexistence of AI and employees. This year alone, the Predictive Maintenance Team attended as many as 10 international conferences and global meetings. Oh said, "Foreign airlines adopted predictive maintenance technology before us, and some have already commercialized their own solutions," adding, "By seeing advanced technology firsthand and engaging with developers, we are enhancing our technological capabilities."