The Japanese government is planning to trial an artificial intelligence system for spotting wild bears to better deal with an increasing number of attacks by the animals on humans, a government source said Monday.
Under the system, AI will use security camera footage to instantly detect bears and share the information with the relevant authorities. The pilot scheme is planned to begin by summer in Toyama Prefecture, central Japan.
If found to be effective, the AI system could be rolled out to prefectures where frequent bear sightings near residential areas have been reported, the source said.
The plan comes in response to the 198 bear attacks on humans across 19 prefectures in the previous fiscal year, which ended in March. The incidents resulted in 219 casualties including six deaths, record highs since fiscal 2006 when comparable data became available.
Scarce food sources for the bears due to poor crops are believed to be behind the increased instances of bears entering human spaces.
The government has stepped up efforts to address bear attacks under a policy compiled in February, and the swift detection of the animals when they show up in urban areas as well as speedy information-sharing among police, municipalities and local hunters have been seen as key.
The new system is expected to be helpful as it connects AI with governmental, municipal and private camera networks.
In the trial in Toyama Prefecture, the system is expected to use cameras installed for surveillance and disaster management by the central and prefectural government facilities, as well as those used by utility companies.
Cooperation from private entities would allow for wider coverage around urban areas, riversides and mountainous areas.
A separate experiment to monitor bear appearances is also planned for Iwate Prefecture in northeastern Japan. It will involve using automatic cameras that can detect animals' movements.
Related coverage:
Gov't panel urges start of subsidized bear culls as attacks rise
Japan mulls designating bear populations for capture amid attacks
