The AI-Driven Female Safety App will be a real-time emergency reporting system, that will be voice initiated and will help support the needs of damaged/unsafe individuals in an unsecure or compromised situation. This solution incorporates Artificial Intelligence (AI) to recognize a list of pre-defined emergency keywords eg; "help," "save me" etc., through an Android smartphones continual background listening ability. This app will be designed to continually listen for these keywords and upon recognition will go through a sequence of automated actions including, haring live location through GPS, recording audio, turning on a siren noise, turning on the flash light, and then automatically notifying the required emergency contacts. This type of tool will allow recordings to operate entirely autonomously and passively - without any action or engagement from the Ontario woman. As above, if a woman is unconscious, paralyzed or restrained, this type of app can be very beneficial. The app will use an efficient, lean model of AI dedicated to keyword spotting that will rely on already established smartphone sensor APIs for microphones, GPS, and flashlight. Not only is the app being designed to detect an emergency accurately and reliably, it will also address concepts of false positives, privacy, and power consumption by utilizing a modular design, with configured configurability thresholds. This step relies on utilizing existing technology that can provide not only personal safety and security (that will become increasingly important for women), but provide a low-cost, scalable, user-friendly safety tool that is using any modern installed version of the Android device and the sensors already included, and not extra hardware from another device.
V P Ilakkiya (Wed,) studied this question.