District in Seoul automates bulky waste disposal with one-shot AI app
An excavator moves discarded small electronic appliances at a recycling facility in Gimpo, Gyeonggi Province, Monday. Yonhap
Pilot program leads municipal innovation by replacing manual forms with automated image processing
A district in Seoul is turning to computer vision to overhaul one of the most tedious chores of urban living: getting rid of old furniture and discarded household appliances.
Eunpyeong District in northwestern Seoul has launched a pilot program using an artificial intelligence (AI) mobile application that allows residents to arrange bulky waste disposal by snapping a single photograph. The service is the first of its kind to be deployed by a municipality in Korea.
Around the world, forward-looking metropolitan centers are racing to integrate machine learning into municipal waste management. Cities like Singapore and Zurich have deployed AI-powered robotic sorters and smart sensors in public bins to optimize collection. However, automating the resident-facing reporting process through consumer computer vision remains a rare frontier in smart city innovation — placing the Seoul district at the forefront of AI-driven urban convenience.
In Korea, disposing of large items requires reporting each piece to the local district office and paying a designated collection fee. Previously, Eunpyeong residents had to manually measure items, look up exact classifications from an extensive municipal ledger and navigate complex online or in-person forms.
The new app, named "Gajyeoga" — meaning "take it away" in Korean — streamlines the entire process into a single step. Powered by advanced image-recognition models, the app identifies the object from a quick camera shot, automatically determines the correct disposal fee and schedules the pickup location and date. Users can complete the transaction and pay without creating an account.
Developed under a public-sector AI initiative backed by Korea’s Ministry of Science and ICT, the National IT Industry Promotion Agency and Kyorim Soft, the system can recognize over 156 waste categories by analyzing an object's contours, material and position — even identifying multiple items in a single frame. Municipal officials estimate the app cuts the time required to submit a disposal request from 20 minutes down to just one.
The upgrade comes as Eunpyeong District handles more than 220,000 bulky waste requests annually, with nearly 29 percent still processed via manual, in-person visits to local community service centers. Discrepancies between self-reported item sizes and actual curb drops frequently caused collection delays or forced residents to re-apply.
Under the new workflow, requests are routed directly to sanitation crews, who can verify items against uploaded photographs on their mobile devices. The app then generates optimized routing paths for collection vehicles and sends automated confirmation alerts to residents once pickups are complete.
The pilot program will run through the end of December, the district office said, while traditional in-person and web filing methods remain available to ensure accessibility for elderly residents and digitally marginalized groups.
This article was published with the assistance of generative AI and edited by The Korea Times.