Korean ministries put AI into care homes, trademarks and boar patrols

South Korea's government is moving artificial intelligence out of strategy documents and into the places where civil servants, nurses and wildlife officers actually work. In a cluster of announcements this month, four ministries each described a different way of attaching the technology to an existing system — not as a pilot for its own sake, but as a change to how a service is delivered.

The most detailed came from the Ministry of Health and Welfare, which reported its "welfare and care AI transformation roadmap" to a cabinet meeting on Oct. 6. The plan's centrepiece is what it calls a "smart home": AI devices combined with home-visit care so that an elderly person living alone is watched around the clock. Today, a crisis is confirmed only after as much as six hours without detected movement; under the new model, AI would combine daily activity, health and conversation data to detect anomalies earlier and connect the person to a service.

In care facilities, night staff now walk rounds every two hours to check safety, sleep and toileting. The roadmap would replace that with smart diapers and fall sensors that detect problems immediately. The ministry said it will develop and test the model in 400 households and 12 facilities in 2027, then spread it through public care from 2028, with home robots offering company, safety monitoring and health management after 2030.

On the administrative side, the ministry wants generative AI to offer welfare counselling in ordinary speech from 2027 and to carry an application through in "one click" by 2029, with universal benefits such as child allowances paid automatically first. It aims to automate more than 50 percent of routine welfare administration by 2028, and to introduce welfare and care "my data" from 2027 so individuals can view and use their own records — later linked to medical data.

Not all of the AI push is about care. The Korean Intellectual Property Office said on Oct. 6 that it had selected 72 new product names in everyday industry use — among them "collaborative robots for unloading," "lidar sensors," "perovskite solar cells" and "immersion cooling equipment for data centres" — and will write them into its official goods classification from Jan. 1, 2027. Filings tied to the three "mega projects" — semiconductors, physical AI and AI data centres — reached 7,016 by August, already above last year's 6,791, while filings for names not yet reflected in the notice neared last year's full-year figure at 142.

At the Ministry of the Interior and Safety, the 14th cross-government startup competition using public data and AI drew 7,440 teams, up about 93 percent from 3,856 last year, for a competition ratio of roughly 744 to one. Ten teams advanced to a November final at KINTEX in Goyang, Gyeonggi Province, where 100 million won ($74,671) in total prize money and ten awards, including two presidential and two prime ministerial prizes, are at stake. The ministry said 82 of 147 past winning teams, or 55.8 percent, went on to found companies, and that 34 teams raised a combined 142 billion won ($106 million).

The Ministry of Climate, Energy and Environment's contribution is more modest in scale but closest to the street. Its National Institute of Biological Resources said it had built an AI monitoring system that uses drones, CCTV and unmanned cameras to detect wild boar in real time and alert officials automatically. A model trained on about 12,000 boar photographs detects the animals in drone and CCTV footage, with drones covering a three-kilometre radius. The institute tested it around Bukhansan, installing one drone, seven CCTV cameras and one alert camera, and will offer the programme and operating instructions to local governments from November. The institute had earlier produced a boar appearance prediction map for the Bukhansan area and gave it to the Seoul city government in February.

The urgency is visible in the call-out figures the ministry cited from the Seoul Metropolitan Fire and Disaster Headquarters: 649 wild boar dispatches in 2023, 589 in 2024 and 494 in 2025, with 236 already by June 2026.

What holds these announcements together is less a shared strategy than a shared method: each ministry is inserting AI into a procedure that already exists — a care visit, a trademark filing, a startup pitch, a wildlife call-out — rather than building a new service beside it. "AI is not a replacement for people but a helper that shares simple, repetitive work," Minister Chung Eung-kyeong said, adding that the goal is to raise the sustainability of care and the value of care work.