Toyo University, Amagasaki City, Hyogo Prefecture, and Fujitsu have developed an AI (artificial intelligence) model that uses contactless sensors such as cameras to estimate whether a person has been deceived by a special fraud.In the future, non-contact sensors will be installed in elderly homes in Amagasaki City to demonstrate the capabilities of the AI model.
According to Toyo University, the AI model measures respiration rate and pulse rate with a camera and millimeter wave sensor based on the data of a demonstration experiment in which you listen to a phone call simulating a special fraud, and determines whether you are being deceived by a special fraud. mechanism for judging.At present, 75% accuracy is guaranteed.
By notifying the elderly of the fraud risk, they can check for themselves what kind of situation they are in.The research team also envisages sending warning notifications to the mobile phones of pre-registered elderly family members.
In April, we installed non-contact sensors in the homes of elderly people in Amagasaki City and started a one-year monitoring experiment.We will conduct a social experiment in Amagasaki City in 4 to verify the ability of the AI model and the effect of reducing special fraud. .
According to the National Police Agency, there were 2022 special frauds recognized nationwide in 3,000, about 1 more than the previous year, and 7,250% of them targeted people aged 86.6 and over.In Amagasaki City, 65 cases of special fraud have been recognized, and the amount of damage has reached about 113 million yen, which is about double that of the previous year.The tricks of special fraud are becoming more complex and sophisticated, and with the progress of the aging society, the establishment of new defensive measures was required.
reference:[Toyo University] Implemented an AI model that uses a non-contact sensor to estimate the deceived state of a special fraud victim