Japan is expanding its use of artificial intelligence to address two distinct public safety challenges: the long-standing crisis of railway suicides and a recent surge in bear incursions into populated areas. In an initiative now active at around 40 stations and commercial buildings, an AI-driven surveillance system designed by Tokyo-based Asilla Inc. analyses live footage from security cameras to identify behaviours indicative of a potential suicide attempt. When it registers a person lingering or pacing anxiously near a platform edge or rooftop, it immediately alerts security personnel and station staff—in some cases also triggering loudspeaker warnings. The technology has already been credited with saving at least two lives, according to reports from Southeast Asian media that have followed the deployment.

Japan has one of the highest suicide rates among industrialised nations, and train stations have long been frequent sites of such tragedies. The system developed by Asilla represents a shift from passive physical barriers to proactive, behaviour-based intervention. By monitoring subtle cues that human observers might miss—particularly during peak hours or overnight—it offers a layer of vigilance that supplements the efforts of beleaguered railway staff. Privacy concerns remain muted so far, given the non-identifying nature of the analysis and the acute public safety imperative.

Meanwhile, in the central prefecture of Gifu, a separate AI-powered tool is being trialled to manage a very different risk. The system, called AIBeS and created by Hokkaido-based Hyke Inc., uses thermal sensors and cameras to detect animals venturing near human dwellings. When its software confirms the presence of a bear, it automatically discharges a commercially available repellent spray. The trial in the city of Hida comes as rural depopulation and changing forest ecosystems have driven a sharp rise in bear encounters across Japan. AIBeS operates around the clock, even in remote mountainous terrain where conventional patrols are impractical.

The parallel rollouts—urban and rural, psychological and ecological—illustrate a broader Japanese instinct to apply technology to social and environmental stressors. Viewed from London or Singapore, Japan’s approach represents a characteristically methodical attempt to engineer solutions that minimise human intervention while maximising precision. European media have noted the bear-deterrent trial as a novel answer to human-wildlife conflict, while Southeast Asian outlets have focused on the suicide-prevention system’s potential as a model for transit networks elsewhere in the region.

As both projects move beyond proof-of-concept, they raise questions about scalability and the algorithmic adjudication of life-and-death decisions. The AI in stations must balance false alarms against the cost of missed detections; the bear repellent must avoid misidentifying children or pets. Yet these early deployments suggest that Japan is willing to embed AI deeply into the fabric of daily safety, blending digital vigilance with human response in a manner that could influence public policy from Seoul to Seattle.