Digital immortality, like ordinary immortality, is a continuum from enduring fame at one end to endless experience and learning at the other, stopping just short of endless life. Preserving and transmitting your ideas is one-way immortality—allowing communication with the future. Endless experience and learning is two-way immortality—allowing you, or at least part of you, to communicate with the future in the sense that the artifact continues to learn and evolve. Current technology can extend corporal life for a few decades. Both one-way and two-way immortality require part of a person to be converted to information (cyberized) and stored in a more durable media. We believe that two-way immortality, where one’s experiences are digitally preserved and which then take on a life of their own, will be possible within this century. We are exploring points along the one-way, two-way spectrum in our CyberAll project (Research. Microsoft.com/~gbell).
Hamarabi, Aristotle, Shakespeare, Mozart, Rembrandt, and Euler are immortal—or at least their ideas are. They recorded their ideas in an enduring form that could be passed on to the future. These great ideas, images, music, writing, architecture, and even algorithms will survive as long as people do. Of course, these people are dead, but their ideas are effectively immortal.
Paper and then the printing press made it easier and less expensive to record, preserve, and disseminate ideas. Voice recorders, cameras, and camcorders now make it easy to record events, and, sometimes, even experiences. Moore’s Law is bringing recording costs down to the point where you can record everything you see and hear.
Digital technologies offer new kinds of information we can convey to the future. They allow almost anyone to create his or her own immortality for any size community—either a family’s future generations or an intellectual community. Web sites (www. 123456789.net, www.legacy.com, www.forevernetwork.com, and www.memorymountain.com) offer (for a fee) to store letters, essays, photos, videos, and stories “forever” in order to pass them on to future generations. These are the digital equivalents of tombs, crypts, and libraries.
Future technologies will surely enhance our ability to convey ideas and experiences, creating a one-way relationship with future generations (should they care to listen or look.) Even today it is becoming reasonable to record everything we read and hear. For example, retaining every conversation a person has ever heard requires less than a terabyte (for adequate quality).
CyberAll is being built along the lines envisioned by Vannevar Bush and Bill Gates as a memory aid and research tool. CyberAll is a store for documents, photos, music, audio, and video recordings, and is currently about 12 gigabytes, including the store for four books, 20 encoded video lectures, 150 music CDs, several thousand documents, and an archive of email messages. It has an accumulation rate of two gigabytes per year. This rate will increase as speech and video become part of CyberAll’s media capture, but it is still a fairly modest expense. Indeed, the real cost of CyberAll is in the data capture, data organization, and data presentation. This is where our research efforts are directed.
Within 510 years, personal stores of a terabyte will cost a few hundred dollars, allowing persons to be immortal in terms of the media they’ve encountered. For “famous” people, one will be able to access his or her entire life.
There are many unresolved technical and social issues associated with CyberAll. How should the information be preserved, given changes in media, platforms, and programs (www.acm.org/ubiquity/views/g_bell_1. html)? How should it be organized and presented? (Will it take a lifetime to see another’s lifetime)? Who should be able to see what, and when? What are the legal and ethical rights and responsibilities concerning information that involves other people? Again, we are exploring some of these issues, but mostly we are focusing on the basic tasks of acquisition, preservation, and recall.
Beyond this one-way immortality, we see hints that at least some aspects of a person could be expressed as a program that interacts with future generations. It is interesting that, given an archive of a person’s spoken output, it is possible to make a compelling avatar of that person. This avatar can “live forever” in a virtual world and respond to queries about that person’s past life. For example, like many great people, Albert Einstein has several posthumous Web sites. In addition, computer science researchers at CMU (cs.cmu.edu/afs/cs.cmu.edu/project/oz/ web/papers/biblography.html) authored an avatar of Einstein that responds to questions from viewers. In fact, the avatar is an actor hired to read quotes from Einstein’s writings. Many who have seen this demonstration understand that in the future it will be easier and easier to author such avatars. The real question is whether such a program could ever “learn” enough to stay current. Having an immortal, interactive program begins to look a bit like two-way immortality—being able to “live and communicate” forever.
We believe along with Ray Kurzweil, Hans Morovec, and others, that it is likely there will be more and more faithful avatars over the next century. By 2040, Morovec predicts robots will be as smart as humans. Successive generations of question-answering avatars will gradually become indistinguishable from the actual persons we know and love in 2001, enabling that person to appear to “live forever.”
We believe that two-way immortality, where one’s experiences are digitally preserved and which then take on a life of their own, will be possible within this century.