I think what is also possible is an avatar of a loved one who has passed. The Avatar would look and sound like them. Through observation of their daily interactions create a model of the syntax and word use maybe even build a history of what they have said where they have gone and what they have done. They could then interact as if doing a zoom call or some future holographic means. The deep fake video and audio exist now, from the demos I have seen on youtube of this research. Probably 5 more years to be completely believable. They are the easiest part. The modeling will require much more to be believable. It is a more advanced problem and has to be captured before the death of the individual. So they have to want this to be made. I don't have any prediction for this or even how many would actually want to use it, although I am sure some widower would like to have it around.