That is unfortunately the way the AI works currently. The recognized objects that are tagged by the AI can be either stationary (parked car) or moving. What is causing the detected motion every minute thought? If your car is stationary, then something else is causing pixel change that is determined by the camera to be motion. The tags are necessary what is causing the motion, its just what’s recognized in the view.
This may be a wishlist that you want to vote on to support.