As a fan of science fiction as presented in a variety of media over the years, I found this topic very interesting. Throughout recorded history, people have used various means to maintain the memory of those they love and admire: first, through oral history, the written word, and depiction of the individual in art, and in more recent history, through photographs, recordings, and video.
Over the past several decades, science fiction has posed the possibility of using the recorded memories and mannerisms of an individual, and applying artificial intelligence (AI) to provide a three-dimensional (3D) construct that appears to act like the original person--the most well-known examples are the holodeck characters on Star Trek. The first steps toward this were in the Disney animatronics, used to provide simple simulations of the 43 US presidents.
This paper discusses this along with more recent work, such as using AI and other information to provide a realistic android simulation of the late science-fiction novelist Philip K. Dick--so much so, that Dick’s family was terribly upset when its creator accidentally left the android’s head on a plane and it became lost. The paper also discusses the ethics of encountering such a simulation and not realizing that it is, in fact, not the real thing--as with ELIZA in the 1970s. Of course, such ethical issues are important.
However, I wish the authors had also addressed the Turing test, that is, whether the simulations can get to a point where they are indistinguishable from a human being and, therefore, considered sentient. This, of course, brings up a whole other set of ethical dilemmas.
In summary, I believe that those who enjoy both science fiction and AI issues might enjoy giving this paper a quick read.