Korean Air inspects aircraft with AI drone swarms and adds a generative AI chatbot
Korean Air is developing swarms of drones and rovers that photograph an aircraft's exterior in about 20 minutes while AI flags possible defects, against more than 10 hours for a manual inspection of the largest types. Since February 2026 a generative AI chatbot has answered customers in 13 languages.
The operating picture
From the hangar to the help desk
Korean Air takes delivery of its first A380 at Toulouse Blagnac International Airport. Photograph: Mattun0211 / CC0 (cropped to fit). Illustrates the airline, not an AI deployment.
Korean Air has its own maintenance, repair and overhaul business, and the drone inspection system is its own development.
It also says technology based on AI improves the efficiency of its fuel management, though it has not described the system.
Selected implementations
The work behind the evidence
01 / Pilot
Drone swarms for exterior inspection
The operational problem
A manual visual inspection of a large aircraft such as a Boeing 747 or A380 takes more than 10 hours, with mechanics working at height.
What the evidence reports
Chan Woo Jung, senior vice president and head of maintenance and engineering, told MRO Management in May 2026 that a swarm of drones, scaled to the aircraft's size, inspects the exterior at the same time while rovers cover the underbelly and engine nacelles. They capture thousands of images, "an AI system analyses this data to identify potential defects", and technicians verify the findings before they enter digital maintenance records. Imaging takes about 20 minutes whatever the aircraft type. Korean Air showed the system at a Seoul technology expo in June 2026, saying it could detect defects down to 1 mm.
Operational significance
Cutting inspection from hours to minutes would shorten maintenance checks and take mechanics off work at height, with a technician still confirming every finding.
Evidence still needed
The sources do not say whether the swarm is in routine use or still in development, and Korean Air has not published accuracy figures for the AI's defect detection.
A source can verify an announcement without verifying its projected outcome.
02 / Deployed
A chatbot that cites its sources
The operational problem
Questions about baggage allowances and fare rules need precise answers, and a chatbot that invents a rule creates complaints.
What the evidence reports
The Korea Times reported, citing Yonhap, that Korean Air launched a generative AI chatbot on its website and app on 4 February 2026. It answers in 13 languages, including Korean, English, Japanese, French, Russian and German, analyses customer intent in real time and handles complex questions such as baggage rules. It was trained on airline regulations and operating policies, uses a specialised database to reduce errors, and includes source citations and links to the relevant pages in its answers.
Operational significance
Citing the airline's own pages in each answer lets customers check what the chatbot says, which matters most on rules and allowances.
Evidence still needed
Korean Air has not published usage, accuracy or handover figures, or named the model behind the chatbot.
Korea Times - Korean Air gen-AI chatbot · Evidence grade B
A source can verify an announcement without verifying its projected outcome.
The wider register
Other tracked activity
Other records are shown separately, including non-AI technology and programmes at different stages of deployment. A register entry does not establish a measured AI benefit.
AI fuel management at Korean Air · Deployed
Korean Air said that AI fuel management helped cut its 2025 carbon emissions by 3.3%, Aviation News reported in February 2026.
A source can verify an announcement without verifying its projected outcome.
Editorial assessment
Korean Air's AI spans maintenance and customer service, and the drone swarm, its own development, is the more distinctive of the two.
Confirmation that the swarm inspects aircraft in routine checks, with accuracy figures, would be the next milestone.
Live data from the Aerostack register. Status reflects published evidence only. See Our approach for how records are graded.
Aerostack register · 3 tracked deployments
Aerostack tracks AI deployments across airports and airlines, with the published evidence behind each record. Photography is real aviation imagery. Record thumbnails may show the named airport or airline; other photographs do not.
