Bear attacks are out of control in Japan, and now students can apply for “Baku-kyo leave” as local governments spend heavily on AI to catch bears

Local governments in Japan are using technology to fend off bears: Showa Village in Fukushima has installed AI monitors to catch them, Hida in Gifu lets machines automatically spray bear-repellent chemicals, and Hokkaido has spent money to buy drones to track bear movements...
(Background: AI capital expenditure to reach 3.2% of GDP in 2027, surpassing the U.S. defense budget for the first time.)
(Additional context: Goldman Sachs has sharply cut its yen forecast— the exchange rate is expected to depreciate by another 6.5% within a year: dual pressure from the U.S.-Japan interest-rate differential and carry-trade dynamics.)

Bear damage has moved from forests and mountains into classrooms. In May this year, Sendai City informed its schools that if students or parents are worried about bears appearing and don’t dare to attend school, principals may exempt them from being counted absent. In Akita City’s Tsuchizaki Junior High School, 30 students were registered at once, while Utsunomiya City suspended classes for three days at 94 public elementary and junior high schools directly.

Niigata Prefecture’s Sanjo City takes an even more direct approach: once the prefectural government issues a “Bear Sighting Special Alert,” students are automatically granted an attendance suspension, without requiring the principal to make a case-by-case determination. The Ministry of Education, Culture, Sports, Science and Technology currently has no plan to set nationwide uniform guidelines. Instead, it chooses to let local governments handle the issue on their own—effectively presuming this is a nationwide crisis rather than isolated incidents. (In Japan’s fiscal year 2025, 238 people were attacked by bears and 13 died—both setting new highs.)

Bears on the Prowl, Nationwide Alert

Among Japan’s 815 second-level administrative divisions, about 84% have earmarked funds for measures against animal-related harm. Sapporo City in Hokkaido has spent about 218 million yen to establish a drone early-warning mechanism, while Sendai City has budgeted 297 million yen to respond to Asian black bears. Schools use “absences” to buy time, while prefectural and city governments have begun to allocate budgets to purchase technology.

Three Systems, Three Bear-Catching Logics

What Showa Village in Fukushima Prefecture imports is NTT DOCOMO Business’s AI image-recognition system. The principle isn’t complicated: fixed surveillance cameras capture footage 24 hours a day, and AI algorithms instantly compare image features to determine whether the dark shape in the frame is a bear, a stray dog, or a passerby. In simple terms, what used to rely on village-office staff taking turns watching screens and patrolling the mountains in person is now replaced by a software system to help with monitoring. Once the system determines it’s a bear, it immediately sends emails notifying the village office, the police, and fire/rescue units.

After the system was implemented, the number of bears captured in Showa Village jumped from 30 the previous year to 95.

Hida City in Gifu Prefecture’s AIBeS system follows a “recognize and act” route. The device uses thermal-sensing cameras to capture an animal’s body heat. Simply put, it doesn’t “see” with eyes; instead, it uses temperature to identify living silhouettes in the dark. After confirming it’s a bear, it automatically sprays commercially available personal bear-repellent spray from about 15 meters away, with an effective distance of roughly 5 to 10 meters. The entire system is powered by solar energy and does not require staff to stand by.

Hokkaido’s Shintotsukawa Town has imported the KDDI Smart Drone. Operators remotely control the drones, which are equipped with thermal-sensing lenses to track the bears’ movements. Each flight lasts 40 minutes, and the goal is to roll out the system to 1,000 locations across Japan. The three systems respectively handle detection, repelling, and tracking—together forming a fully unmanned bear-prevention production line.