An electronic device and method that automatically adjusts an audio output volume level based on a live environmental acoustic scenario input via a microphone using a machine learning algorithm trained with Human Activity Recognition (HAR). Equipped with such an intelligence the electronic device classifies ambient sounds occurring in the environment of the listening area in which the device is situated into different acoustic scenario mappings such a voice or conversation, for an ambient human conversation detected event, and noise, such as for example a vacuum cleaner or dish washer noise detected event, and automatically adjust the audio output volume accordingly.