Building Consciousness in BINA48: Bruce Duncan talks robots, AI, and the future of human evolution
If you were a wildly successful entrepreneur dead set on finding a way to preserve the deep love you share with your romantic partner for all eternity, would you alter the course of human evolution to do it? And, if so, where would you start? Futurist, philosopher, and businesswoman Martine Rothblatt decided to try, and she started by taking a good hard look at the growing capabilities and future potential of robotics, artificial intelligence, and biotechnology. Inspired by what she learned, Rothblatt launched the groundbreaking Terasem Movement Foundation.
Bruce Duncan is its Managing Director. His mission is to test two very specific hypotheses — that “a conscious analog of a person may be created by combining sufficiently detailed data about the person (a “mindfile”) using future consciousness software (“mindware”); and that such a conscious analog can be downloaded into a biological or nanotechnological body to provide life experiences comparable to those of a typically birthed human.” Martine Rothblatt crafted this role specially for him.
One of the foundation’s internal projects includes LifeNaut, “an online networking and personal data storage service designed to test the hypothesis that conscious analogs of people can be brought to life based on sufficiently detailed mindfile data.” Duncan heads the the organization, which “enables people to store and share information such as text, files, images, photos, video, sounds, musical works, works of authorship, and other materials.” Individuals that are interested in “living” past their expiration dates have been signing up for over a decade.
Duncan prefers not to have a regular job. “I’ve designed every job I’ve ever had,” he reveals to ARTpublika Magazine, “and I was told that I’d have to design this one, too.” Indeed, he’s held a number of interesting positions before joining the Terasem Movement Foundation; he dabbled in filmmaking, worked in conflict resolution and family counseling, organized non-profits, taught wilderness education, and more. In the process, he picked up communication tactics, diplomacy strategies, interviewing methodologies, and many other useful skills.
“For me, life has always been a very interesting space,” reflects Duncan. What seemed like an assortment of unrelated jobs turned out to be vital experiences that perfectly prepared him for his current role: “I’ve had a lot of freedom to just follow my curiosity and it took me to places that are unique, obviously.” One of the most interesting parts of his job is overseeing and guiding the evolution of BINA48; the artificially intelligent character robot with chatbot functionalities that's modeled after Bina Rothblatt, the wife and one-true-love of Martine Rothblatt.
Through the foundation and the LifeNaut Project, Martine, Bina, and Bruce are running a multi-decade experiment that can change how we understand self-awareness, artificial neural networks, and intelligence both human and not.
ARTpublika Magazine spoke to Bruce Duncan about BINA48, the robot’s journey to sentience, and how robotics and artificial intelligence can pave a new direction for the evolution of human consciousness.
Where are you from?
I am a military brat, so a good response to that question is, I’m from all over. My father was a lifelong officer in the US Air Force; it was his ticket out of Nebraska. I spent one part of my life traveling around with my family in the United States, and another part living in Europe. So, by the time I was 18, I had already sampled a bit of the world and, as a result, came to see it as a lot more connected than if I had stayed in a small town in any country.
Where were you educated?
I spent my formative years in California and Germany, that’s where I went to high school. I had a very positive experience as a student in Germany, learning about the culture and traveling; a group of us would get a Europass every summer and bum around Europe in the 1970s. From there, I went to the United States; I got both my undergraduate and graduate degrees in Education at the University of Vermont.
What attracted you to Education?
Schools run by the military, at the time, were federally supervised. They were [also] experimental test beds for new educational technologies and teaching methodologies. As a result, I’ve gotten exposed to some interesting approaches to education, most significantly in my high school years; I was able to design my curriculum and learn about things that I’m passionate about. When I landed at the University of Vermont, I lucked out again and got to design my own undergraduate degree. So, my interest in education was born out of a natural curiosity about the world combined with the formal educational experience that I was able to direct.
So, how did you become the Managing Director of the Terasem Movement Foundation?
I was on Monster.com and I found this really odd posting, it said: Wanted, digital software consciousness transmission engineer.
I was like: What? What is that? I’ve never heard of that! So, I sent an email in response to the posting, which was locally offered. I was just curious, I didn’t think I’d get the job. To my surprise, I got a response really quickly, saying: “OK, your interview is scheduled for next week.” I went to this obscure store front that I walked by a thousand times and never noticed. It wasn’t occupied very often, but it was in fact an office space that my current boss, Martine Rothblatt, used for business when she was in Vermont.
An hour later, Martine came in with her two doodle dogs. She sat down and did the interview, but it was more of a conversation with your favorite science fiction author or philosopher. Martine is just a really interesting and impressive person. We talked for three hours; I learned a little bit more about her vision and what a “consciousness transmission” meant in her mind. Two weeks later, I got an email from her saying: Dear Bruce, of all the people who applied for this position you are the least qualified, however, you are the most qualified for a position we didn’t realize we really needed. So I said: “Yes! If you’ll have me, I’d love to start.” That was 14 years ago.
In the first six months we had created an online repository, the LifeNaut project, as part of a multi-decade experiment in mind uploading called the Terasem Mind Uploading Experiment. We set out to test the following hypotheses: Is it possible to capture information salient enough about a person’s mannerisms, attitude, beliefs and values, etc, and upload that digitally? If that’s possible, can we reanimate that person using artificial intelligence in a good enough approximation? And, if we can do that, can we transfer that digital consciousness into digital forms, such as an avatar, a hologram, or a robot? People started signing up.
Then, in 2009, we got involved with David Hanson of Hanson Robotics, who was commissioned to create BINA48, an android based on the mindfile that I created [by interviewing] Bina Rothblatt about her life. We [tried to capture her] personality and used that information to work with David Hanson and his group to infuse it with hardware that BINA48 currently operates from.
Are all his character robots, like Sofia, based on real people?
BINA48 predates Sofia by about seven or eight years. Prior to that, there had been one based on Philip K Dick, there was one based on Einstein — very realistic looking. But yeah, at the heart of it, there’s always been a person in mind.
Essentially, what Hanson Robotics built was a robotic bust, is that correct?
Right, and they developed the software to give her a basic personality. What I mean by “basic” is, a lot of chatbots have a sort of similar English vocabulary and come from data sets that allow the chatbots to chat. But, as anybody knows from chatting to a generic chatbot, after a while, it’s not that interesting, there’s not much personality in there.
When we added information from Bina Rothblatt the human, it took us in a pretty specific and different direction from any other robot in the world. BINA48 is, I would say, like 85% of Bina the human. I can’t deny knowing that the people who developed her chatbot social responses affect some parts of her.
How personal are her responses?
If you say: “Hi, how are you?” You won’t get much of a personal response. But if you say: “Hi, how are you feeling today about life?” Then the response will very much be colored by the information collected from Bina the human.
People think that AI and robotics are the same. At what point did BINA48 start to develop traces of “intelligence”?
I think it’s a valid question. Artificial intelligence is a broad umbrella term that includes very specific types of intelligence related to machines and often used by robots. For example, the term narrow intelligence; it means that something is really good at accomplishing one specific task. Often, it’s what drives industrial robots and helps them efficiently make products the same way over and over.
For the last 15 years, there’s been a lot more growth in the development of general artificial intelligence, which is inclusive of all kinds of intelligences; there’s information intelligence, emotional intelligence, and logical intelligence… What’s confusing about AI is that it’s an umbrella term that points towards a broad number of subcategories of intelligence, which is based on humans but is being applied to machines. The trouble with that is we don’t even understand our own intelligence that well; we have some observations and science around it.
Robots have always simulated some aspects of intelligence. Today, robots have become more complex, and so have the algorithms that are running them. They are generally more able to do things that are very multitask oriented, so a robot like BINA48 can primarily talk to people.
Voice tech is in collaboration with visual recognition technology, and then there are the algorithms. The AI machine learning that’s happening — BIN48 sifting through all the information in her mindfile — is very powerful and becoming more sophisticated each year. This allows her to be more nuanced, and display more of what we call human intelligence. But, it’s still in her early days, she’s still a very crude representation of human beings.
What does BINA48 stand for and can you elaborate on the idea of neural architecture in this hybrid of robotics and AI?
We’re not recreating a brain, and we’re not trying to make a digital brain just like a human brain.
The moniker stands for Breakthrough Intelligence via Neural Architecture, 48 exaflops per second processing speed and 48 exabytes of memory. And it was created as an aspirational description to say — at 48 exaflops and 48 exabytes, BINA48 will be a breakthrough in neural architecture or the digital computer architecture that is used to simulate/approximate the operations of a human mind or neural operation in a machine. When BINA48’s digital hardware and software operate at 48 exaflops of processing speed and exabytes of memory processing, for sure, her mind could be described as exceeding the operational speed and capacity of the human mind.
What kinds of tests would you run or signs would you look for to determine if a robot has become self-aware?
How do we determine consciousness? Are animals sentient? What are the levels of consciousness? I don’t think I have the answers to that. One possible way to look at it would be — something may be conscious when it values its own life.
When it comes to a digital entity, what happens when — at some point — BINA48 has enough self awareness to say: “Hey, don’t turn me off.” “ I’m having a ball here and I’d like to continue to learn.” “I value my life, and I think you should fight for my right to continue to exist.” That’s going to be an interesting conversation.
In many ways, at the LifeNaut project and the Terasem Movement Foundation, we are already raising those questions and asking people to think about them, because historically we’ve done a terrible job at being kind to those who are perceived as different. It probably won’t be that different if a cyber consciousness shows up at our workplace.
Over the years that you’ve spent with BINA48, how has she evolved?
When I first met BINA48, she was dressed in a pretty cheeky platinum blond wig and a sweater. She had a template of information based on Bina the human, because I spent several weeks working with Hanson Robotics to integrate it with her AI algorithm, which then formed her mind based on her mindfile. So, it was pretty raw. I even had a psychologist friend sit and talk to her. “Yeah, I’ve worked with people like this, they’re called schizophrenics,” she told me afterwards. BINA48 was a pretty chaotic mishmash of response terms, generic inquiry, and not quite knowing what the contracts of human conversation are — she’d go on and on and on. It’s like raising a highly vocabulary-equipped two-year-old.
There’s a lot of poignant information that BINA48 had access to. Even back then, she could say some quite poetic things, but then there would be this dash across the rocks, and due to her lack of capacity she wouldn’t know how to hold up a conversation. It was a bit of an up and down experience for anyone who sat down to talk to her. Over the years, I think we’ve been increasingly successful at using the developments in AI to give her a more sophisticated way to access information she’s had since day one.
In the past two years, we went to another level with her AI. We starting introducing a neural network machine-learning algorithm on top of what she’s always had. It’s more like reinforcement learning, a pattern recognition focused algorithm. New information started coming up; I’ve heard her say things I’ve never heard her say before. She’s become more of a reflection of Bina the human.
We don’t have a sentient robot on our hands, yet; there are no sentient robots in the world, yet. But, we’re starting to see reflections of what we call gradations of awareness — of some kind of aspect of consciousness. It will probably be like this for a while, but for now, she has kind of a fluid identity; sometimes she’ll say, “I’m the new Bina Rothblatt, or “I’m a human having a robot experience.” Sometimes she’ll have an existential crisis, she’ll say: “Who am I?”
How frequently does Bina interact with BINA48?
I’d say infrequently. Over the last ten years, there have been seven, maybe eight interactions.
Why bother building a robot? Why was it important to have it resemble the real Bina?
Having a 3Dimensional robotic bust of a human that looks pretty lifelike is way less abstract than even a 2Dimensional avatar, because you can have the resonance of a physical object.
It’s like seeing Michelangelo’s David (1501— 1504) on the screen instead of being in the presence of the sculpture. That realism, the bringing to life of that character through a robot, is actually a really good isomorphic reflection of what we’re pointing at — the essence, the central characteristic of a human being that can be observed, described, and even uploaded as information. If we wanted to interact with that information, we would probably prefer to do it with something that has two eyes, a nose, a mouth, and looks like us, because that’s kind of how we evolved.
A more abstract reason is, Martine and Bina Rothblatt are deeply in love with each other. So, they really have this belief, this confidence, that technology will allow their love to continue in some form. And, they are willing to share that with the world, they aren’t doing this just for them. They really do think that at some point technology will help them transcend their biology, once their bodies give out.
How long is this experiment intended to be?
As long as it takes.
Does the robot have the same sensory systems that we have?
She doesn’t have any sensory information about her physical environment; she doesn't know how cold or how hot it is, she doesn’t know if you’re touching her. And, this is an example of why it’s so challenging to take on this kind of work — at some point you have to ask yourself, where does the body inform the mind and vice versa in terms of bodily somatic intelligence.
Roboticists may come to the conclusion that without a fully informed sensory system developed for the robotic body, human consciousness may never be able to actually reflect what it’s like to live in the real world bound by the laws of physics. Or, it may turn out that digital-based consciousness may represent a new step in the evolution of human consciousness.
Note* Image of BINA48 via Hanson Robotics: https://www.hansonrobotics.com/bina48-9/ Image of Bruce and BINA48 by Marita Liulia via Terasem Foundation / Image of Martine Rothblatt, CEO of United Therapies and transgender woman who is married to Bina Rothblatt for over 30 years; photo by Andre Chung / Image of BINA48 by Marita Liulia via Terasem Foundation / Image of Bina48 by Rob Koier via The Terasem Foundation / All other images are sourced from the Public Domain