I agree that this is the most interesting part of the article, sociologically. My father was a tech guy who passed away over ten years ago, I believe that if you had the power to do so it would be an invaluable gift to your loved ones to leave them little messages every so often… (as long as your relationship was healthy!) But unlike the attached article implies, psychologists shouldn’t be concerned as I don’t think most people will ever feel like we are truly “communicating” with our deceased loved ones as communication is always a two-way streak.
A robot giving standard-human like answers (or a relative enabling that) isn’t really your loved one communicating unless that loved one somehow took the time to sit there and guess what you might answer in different scenarios. As much as I’d love periodic messages and jokes sent from my dead father to appear every once and a while (and would look forward to and treasure them) I could never feel compelled to reply back to him, or desire a robotic answer, and once you cross that line then I think it falls into “socially unacceptable”.
What about when we get to the point where we have an AI capable of simulating, if imperfectly, an individual's response to a given question?
If you think that this is farfetched, it's not - it's basically like http://iwl.me , but in reverse. The catch is that you need a lot of easily classified textual data about the person. But in the Twitter age, we do have that kind of information - or soon will.
Gracious Eloise[1] is a really cool idea, though it's going to create a nightmarish situation for forensics experts. Imagine if you combined the two - you could simulate what a person might say, and you could simulate their actual handwriting for it.
