I have a camera is set up in the room of a very young cancer patient who is dying to make sure they are being cared for appropriately by hospice and facility nursing staff.
 

I had planned to share it with other relatives to make it easier to communicate with other caregivers and was horrified to hear the cutesy audio overlay. It’s wildly inappropriate for our use case. There should be a setting to disable adding any audio that wasn't recorded on the device and an option to re-record old summaries when that setting is toggled off. As it currently exists this function is useless and I'm extremely disappointed that more thought hasn't been put into this feature.
 

Additionally, it doesn't appear that the summary stitches together ALL of the individual recordings which makes this feature useless for monitoring a dying hospice patient. The settings for this feature should be much more robust. 
 

Given what AI is capable of doing it would also be incredibly helpful to get a timestamped text summary of her day.
 

Between the video recordings and the audio that was recorded it shouldn't be difficult to generate a text file that says at 8:17 a diaper was changed & at 21:00 she was repositioned & she was offered food/water at X time. 

This may require a long text field in the setup workflow to allow users to enter a prompt of what the camera is being used for and what output is expected from a text summary. 
 

For example, ours would say, "This camera is being used to record the activity in a hospice patient's named Jane's room. Nursing staff usually wear blue or green scrubs. Critical information to monitor are the time & frequency of Jane's diaper being changed, the timestamp of how frequently she is repositioned in bed, how often she is being offered food or liquids. Additionally we would like a summary of any non-staff visitors she may have. We would also like to know if staff is ensuring that there is some sort of music or television being turned on during the day to provide mental health stimulation."