Provided are an intelligent interaction device and method, and a non-transitory computer readable storage medium. The method includes: displaying, on a touch screen, a current window of a multimedia file in a playing state; displaying, in response to an instruction from a user for zooming the current window, a zoomed window of the current window at a first predetermined position of the current window, the zoomed window being smaller than the current window; and displaying, in response to an annotation operation performed by the user for the zoomed window, an annotation in the zoomed window, and updating the current window by displaying the annotation in the current window.