Welcome to your future android clone
March 12, 2012 by Amara D. Angelica
Bina48, first android? (credit: Terasem Movement Foundation)
This is the most interesting event at SXSW I’ve heard of so far: “Robot panelists, AI and the future of identity.”
It’s a session Monday at SXSW (3:30PM – 4:30PM), where Bruce Duncan, Managing Director of the Terasem Movement Foundation, will bring us up to date on Terasem’s amazing LifeNaut project.
LifeNaut is a free online service (and experiment) for personal data storage and avatar interactivity, says Duncan. “It allows people to build a rich personal profile that preserves their essential, unique qualities as ‘mindfiles.’” (They also allow you to store your DNA.)
“Mindfiles are database files with uploaded digital information (videos, pictures, documents, and audio recordings) about a person’s unique characteristics (such as mannerisms, attitudes, values, and beliefs),” he explains.
Mindfiles (there are about 12,000 so far) are stored online at lifenaut.com. Future AI programs, Duncan believes, will use a mindfile and a person’s DNA to create a digital clone of that person that can interact with future family members and others.
As an experiment, the Terasem Movement Foundation has created Bina48, an android based on the mindfile of Bina Rothblatt, cofounder of the Terasem Movement Foundation. (Full disclosure: I introduced Terasem to Hanson Robotics to build Bina48.) Bina48 has gigabytes of information uploaded to a database. “She” continues to acquire new experiences and knowledge by interacting with people (using vidcams in eyes, face-recognition software, and Dragon voice-recognition software).
(Ray Kurzweil has developed a similar avatar called Ramona that exists as a software chatbot on KurzweilAI.)
Creating an intelligent clone
Taking it a step further, AI researcher Stephen L. Reed will describe how his Texai AGI (artificial general intelligence) system could (in theory) one day be used with Lifenaut mindfiles to emulate human intelligence. Reed was formerly a project manager at the famed Cycorp, where he developed OpenCyc.
“Texai’s software uses a dialog to acquire facts and incorporates them into a sort of knowledge base,” he explains. “Texai also uses natural language processing, as in Siri, except that Siri is only used to contextually disambiguate a user’s query or utterance, to plug into a particular iPhone service, Wolfram Alpha, or web search.”
In contrast, Reed created a lexical (dictionary-like) knowledge base by merging WordNet, Wiktionary, and OpenCyc software. “From the latter, Texai has mapped about 12,000 noun-word senses to OpenCyc semantic concepts. WordNet and OpenCyc provide taxonomic relationships between these terms. The system will seek to map additional word senses to noun, verb, adjective and adverb concepts, in a semi-automated fashion, via dialog with an expert user.
“One could imagine that LifeNaut might seek to preserve a user’s skills, so that the avatar could ‘do stuff’ even after the user is deceased. If those skills included the skill of learning, the avatar could subsequently improve itself.”
Reed also intends to pursue the method Alan Turing recommended in 1950: “Instead of trying to produce a programme to simulate the adult mind, why not rather try to produce one which simulates the child’s?” said Turing. “If this were then subjected to an appropriate course of education, one would obtain the adult brain.”
Comments (19)
by GatorALLin
…what about David…. version 8 is now here..
http://www.projectprometheus.com/
by John Doe
This thing might probably be capable of simulating me if I were to have a lobotomy followed by a stroke.
by anonymous
hahahahahaha AGREED!!!!!
by anonymous
did they put the face through a tanning bed? it’s a bit scary
by eric25001
Why not?
Why not replace all Federal, State, and Local government workrs and contractors as they reire? Provide sat at home bots for cooking, cleaning, child care, elder care, emergancy first aid?
Will work become something people use to do?
What if bots become as or more creative and inventive as humans?
Can a bot go to space? Take human DNA to expand the human domain?
by cosmowrench
I dont see why you should upload a “mindfile” and trust a comany with all that info. You can just as well put one together yourself and keep it on your desktop.
by TheCyberLance
How can I do that? I want an avatar that looks like me and can speak on my desktop to my family when I’m gone. I have been looking for two years for something to do this and have yet to find anything. If you know how, please share it. For now, Lifenaut does what I want and I’m happy with it.
by Cybernettr
LOL Seeing that head and shoulders person reminds me of the character Admiral Pike from the original Star Trek.
by Anonymous
In reply to
Bruce Duncan
Managing Director of Terasem Movement Foundation Inc.
First, as to the writing employed in the agreement (as one example only), any lawyer who writes a clause in the form of “(1) blah, blah, blah; (2) blah, blah, blah; OR (3) blah, blah, blah AND (4) blah, blah, blah” not only missed out on contracts class but high school freshman grammar as well (which goes to say that litigation is invited over whether all of the subclauses are alternatives to each other or combinations or just (2) and (3) are alternatives or whether (3) and (4) go together or are separate or what goes with what or what exactly the clause means – and, by the way, unintelligible contracts get construed against the drafter). Astonishingly poor.
Second – and more to the point – my comments were directed to this website – not yours – regarding their policy of pointing people to websites where they may unwittingly give away their life stories without so much as a mention of warning in that regard. While it is up to the individual – ultimately – to be responsible for themselves, a referring website that prides itself on its reputation should realize that that reputation will cause others to act simply because of something it says – and the website should act accordingly. I certainly do that in everything I do.
Of course, the fact that kurzweilai.net allowed you to reply rather than address the matter themselves only goes to show the degree to which your organizations are apparently in bed with each, without bothering to mention that in the original article. Quite disappointing over all and something that knocks my opinion of this site down more than a notch for me since now I will have to question the degree of interconnectedness and bias about everything written here.
by Bruce
I agree with the suggestion that people should really read the Terms of Use agreement before signing up for anything on net and yes lawyers were involved in creating this TOU like most TOU’s. It is meant to be a legal agreement between people who sign up and the LifeNaut project. In the case of LifeNaut we wanted to make it clear that when people sign up for a LifeNaut account that they are doing so in support of the research project we are conducting into the possibility of capturing the essential characteristics of person and maybe even one day downloading that information into “new” forms or analogs one day. This is an important TOU detail because as non-profit we are not selling LIfeNaut accounts as a service, although there are benefits to people who sign up e.g. creating a free mindfile. We mentioned the ownership details because currently that is an emerging issue right now where most social networking sites own your information for their use with varying levels obligation to what happens after you pass. So in addition to making your wishes know about how you would like your LifeNaut account treated in the event of your passing, LIfeNaut by putting it in your will, LifeNaut will soon be implementing a “proxy” option that allows you to name a person who will be able to choose to continue or delete your account for you. My suggestion is to think carefully and make sure you understand the LifeNaut TOU agreement before signing up. If you have any questions please contact me at terasem@gmavt.net.
Bruce Duncan
Managing Director
Terasem Movement Foundation Inc.
by Anonymous
I went to the Lifenaut site to poke around with the idea of perhaps opening an “account” there. Then I took a look at their user agreement. It reads like a lawyer wanna-be wrote it and includes such niceties as they own all of the information you provide if you die without a provision in a will. Or that the site can use the information to create “analogs” in the future. And it isn’t at all clear that you can object from what I skimmed through.
No, thanks.
In the future, you should consider such user agreements before writing an article since I’m sure many people will sign up simply because an article appeared on your website.
by TheCyberLance
I read the agreement before I started my mindfile and I have no problem with it. I have my reasons for making one and it’s my choice to do so. Maybe if you were in my situation you would understand why I feel this way and be more understanding of this technology.
by William Hertling
The Bina48 demonstration at SXSW was interesting (I was in the audience), but the real point of the project is not to create a chat-bot, but to recreate a specific person’s personality.
The downside of the demo is that while it is interesting, none of the attendees known Bina Aspen, whom Bina48′s personality is based on. So while Bruce (from Lifenaut, who brought Bina48) knows Bina Aspen, and can see the similarities to the original person, that’s impossible for us in the audience to see.
Obviously it’s difficult to onboard all the necessary data, but what will make a more convincing demo is when there’s a Bina48 equivalent for someone with a well-known personality, so that many more people can experience the effect of recognizing aspects of that person’s personality.
Here are some of my session notes from their panel: http://www.williamhertling.com/2012/03/robot-panelists-ai-and-future-of.html
Thanks,
Will
by Editor
Will: thanks very much for those very helpful notes for both panels. I plan to read your very relevant science-fiction novels. – Amara
by egore
Hey Ramona, welcome to the world of dingalings!!
by Spikosauropod
Reanimate a person’s consciousness?
I consider such light treatment of consciousness to be very dangerous. If people think they can “reanimate” consciousness with a file, they are likely also to think the original consciousness is expendable.
Here is some required reading:
http://consc.net/online
You should keep reading papers on this site until you have either exhausted the list or you comprehend the problem.
by Khannea Suntzu
Conversation with Binah herself is somewhat more uplifting. :)
by Bruce
Its true that Bina48 is pre-programmed, but that misses the point. She is not meant to demonstrate a robot that is fully “awake” and independent but rather a first step towards that goal, by showing how information from a mindfile about a specific person can begin to represent that person through current our technology. We are interested in the question: that given a robust database saturated with the salient information about a person will future AI software be able to reanimate that persons consciousness. This is just the first step in our effort to pursue that question.
by Anonymous
She’s a glorified chatbot, most of her responses are preprogrammed.