AI-Generated Bear Videos Fuel Panic Amid Record Fatal Attacks in Japan

Amid Japan’s record 13 fatal bear attacks in 2025, AI-generated videos showing fake encounters on platforms like TikTok are spreading misinformation, heightening public fear and complicating safety messaging.

By Maria Konash Published: Updated:

Japan is facing an alarming surge in bear attacks, with 13 fatalities and over 100 injuries reported this year—more than double previous highs. Experts attribute the increase to food shortages caused by poor acorn and beechnut crops, and the encroachment of human settlements into bear habitats.

Compounding public anxiety are AI-generated videos circulating on TikTok and other social platforms. Many depict bears attacking people, entering buildings, or interacting with humans in unrealistic ways. A Yomiuri Shimbun investigation found about 60% of sampled clips were AI-generated using OpenAI’s Sora technology, often including fake news-style reporting and realistic location details.

Authorities warn that such videos undermine official safety guidance, including warnings not to feed bears or confront them, which can be deadly. Experts stress that black bears can weigh up to 130 kg (287 lbs) and Ussuri brown bears as much as 400 kg, easily outrunning humans.

The bear problem is particularly severe in northern Japan, where increased sightings—20,700 between April and September, up 7,000 from 2024—have disrupted daily life. Japan Post, NTT East, and Yamato have all adjusted operations to protect workers, and self-defense forces have been deployed to assist in trapping bears.

Authorities urge residents to rely only on official information, ignore unverified videos, and follow established safety protocols to avoid dangerous encounters.

AI & Machine Learning, Consumer Tech, News