Lots of people who consider themselves futurists expect a day will come when people will be able to transfer their consciousness to non-biological entities—uploading their minds into whatever the post-singularity version of a computer is. Leaving aside whether "the singularity" as a concept makes any sense (hint: it doesn't), what might this transfer process actually look like? Would it involve scanning the brain with some kind of imager, or hooking up wires to specific locations in the nervous system, or something else entirely?
Recently Ars was made aware of a group out there that claims to be actually preparing for the happy day when human consciousness can inhabit a machine. In fact, it's already experimenting a bit with some initial attempts at making that happen. And the results are thoroughly bizarre.
The ideas behind the effort come from an outfit calling itself the Terasem Movement Foundation, or TMF. To be frank, the organization comes across to us as a mash-up of a cult and the kind people who are supposed to be keeping Walt Disney's corpse on ice. The group's Website is littered with jargon-laced text like this: "TMF safekeeps mindfiles and biofiles of lifenauts for future revitalization in accordance with their consents and technology advancements, while also spreading the good word that software people are people too."
The group seems to think that it's possible to capture the essence of someone's personality and store it in a computerized form. When the technology is eventually ready, that stored form can then be used to create an artificial intelligence shaped by the original person's personality. Willing "lifenauts" who can't wait to get a jump on the robo-future can start the process now. The proof-of-concept "robot" here is named BINA48, and the bot's appearance and "personality" is based on TMF founder Martine A. Rothblatt's spouse, Bina Aspen.
If any of this group's ideas held water, then you'd think we would have seen a bit of excited buzz as the preliminary work in the field was published in the scientific literature. This hasn't happened, but hey, the website's got a science page, so the foundation must at least be solid, right?
No, of course they're not. Or, rather, they're an interesting mix of good stuff and weird misinterpretation. For example, the page on neural implants correctly recognizes that such implants can "be used to record brain system functions with unprecedented detail." But it also suggests that microchip implants can "repair damaged or diseased neurons"—and at least for now, there's not even a hint that this will ever be possible.
The section on biology ascribes all aging to a single enzyme and warps right into the realm of science fiction with this: "It is not unlikely that engineered human bodies may be grown in the future. Together with nanotechnology, these bodies may be engineered to accept the mindfiles people had stored back in 2007 at lifenaut.com!"
So, we're dealing with an organization that has an understanding of science that ranges from the optimistic to the iffy. Which is a bit reassuring, given that the group has managed to create a chatbot with an odd fixation on the Terminator films. At least if the chatbot manages to somehow gain sentience, it probably won't have the tech to create advanced weapons right at hand.
Regardless of the ideas behind it, the chatbot "AI" in the videos appears capable of an interesting level of interactivity. And the website suggests that BINA48 was built using a mixture of surveys that create a personality inventory, and along with automated scanning of video and first-person writing. The site even discusses (again, probably optimistically) building similar profiles simply through social media material.
We are skeptical that this would really work. Personality inventories are meant to provide general tendencies, not an actual reconstruction of the person. For example, an inventory can identify things like openness, or an interest in new experiences. Openness then influences things ranging from ideology (conservatives, not surprisingly, tend to be less open) to interest in travel.
But these are just tendencies. There are definitely open conservatives out there, for example. There are also people who might be open to trying new food, but have allergies or a bad experience with food poisoning that keeps them from following that interest. Can all of these sorts of details be captured through interviews and blog posts? Our guess is no. The process produced something interesting in the BINA48 bot, but we would be surprised if it fully captures the personality of an individual.
There's probably something interesting to be learned from whatever this system is. But we'd need to have a full description of it to have it go beyond being a curiosity.
reader comments
132