Perfect recreation? Then according to my philosophical beliefs, my loved one is not actually dead and we can share a laugh about a month spent doing neural pattern scans at the hospital after a little accident before I was able to talk to them again.
Basic information retrieval system for my archive of photos / videos / e-mails, like Copilot is an information retrieval system for the web? Would not replace being with the person of course, but certainly helpful to reminisce about good times and clarify details relevant to the living.
Something in between? If both ends of the spectrum are good, why would the middle necessarily be bad? ML can leverage knowledge of myself contained in the archives to give me better life advice than a friend or therapist without such extensive knowledge would be able to.
There is of course abuse and malpractice potential for companion AI, but obvious potential good far outweighs the risks.
15 comments
[ 2.2 ms ] story [ 47.3 ms ] threadMark Zuckerberg was on Lex Fridman and talked about this exact type of thing
I hope this is embraced rather than pushed away
The AI is not going to achieve that, only a fake version of it.
But, you know, I treasure the photographs of my late wife and wouldn't part with them for the world.
I was reminded of this episode. From this episode it felt as though she was prevented from letting go and moving on by the AI.
https://www.youtube.com/watch?v=EYvrziE4feI&t=70s
Let the dead rest, they’ve earned it.
Basic information retrieval system for my archive of photos / videos / e-mails, like Copilot is an information retrieval system for the web? Would not replace being with the person of course, but certainly helpful to reminisce about good times and clarify details relevant to the living.
Something in between? If both ends of the spectrum are good, why would the middle necessarily be bad? ML can leverage knowledge of myself contained in the archives to give me better life advice than a friend or therapist without such extensive knowledge would be able to.
There is of course abuse and malpractice potential for companion AI, but obvious potential good far outweighs the risks.