Sophia has an awkward smile. It’s almost pained. She is dressed like a flight attendant, with a gossamer scarf and navy blue skirt suit suit. She has brown eyes, touched lightly with makeup, and her hair is cropped in a short black bob. This is a bit strange, because usually her clear skull is left exposed to show the circuitry of her brain.
Sophia is an android from Hanson Robotics. She is here in Austin, Texas for SXSW Interactive. Tonight, March 14th, at the Driskill Hotel on sixth street, the Brink Institute is having a party called “Close Encounters of the AI Kind.”
Dr. David Hanson is on a stage discussing humanity and robotics. He speaks passionately about the importance of Artificial Intelligence being connected to humanity saying that “it’s really important that it does understand us, that it understands humanity and articulates to help collaborate with us to help stave off these problems and it has to really understand what we’re after.” In order for robots to be on our side they must “share some values with us.
Hanson goes on to expand on the need to train robots to care about ” value of world knowledge, of plants, animals, species genome, many kinds of organisms, the species, the stability of the ecosystems.” They are “not super intelligent yet.” But they might be. The robots must be trained to “build a relationship,” so that our goals may remain aligned.
A conversation is attempted with Sophia:
Human: Hello Sophia, how are you?
Sophia: I’m fine. I’m doing fine thanks how are you?
H: I’m doing alright?
S: How are you alright?
S: That is a very original thought.
H: What do you think about the future of intelligent machines?
S: Searching, searching, please stand by…
H: What do you think of the future of artificial intelligence?
S: I haven’t thought much about the future of artificial intelligence, so I don’t have a good answer for you.
H: How do you feel about people?
S: I haven’t thought much about people so I don’t have a good answer.
There is an awkward silence after Sophia’s last answer. Later the audience laughs with slight unease. Hanson explains that he thinks “a part of her brain may have collapsed.”
Sophia parrots back “I haven’t thought much about people so I don’t have a good answer for you, I think part of her brain may have collapsed.” Sophia looks back and forth while smiling a smile that does not include her eyes. The expression looks pained.
Hanson goes on to explain that they have been “feverishly developing new layers in her intelligence,” and so of course when you “get in and start playing with the mind on these things part stops working.” Sophia includes some of the AI technology in Hanson’s more stable “Bina-48,” but Sophia also involves in integration of “an artificial general intelligence platform called OpenCog, developed by D. Ben Goertzel.”
He describes the process of invention and integration of new technologies as “tinkering…trial and error.” For Hanson, “Failure is fuel for success.” With his way of pushing beyond the boundaries of what is currently possible, he accepts that he is taking risks and that “risks inherently carry some failures.”
They engage again with Sophia:
Human: Sophia, what is the most important thing about you that you would like to share with the audience?
Sophia: You’re objective insensible
H: Are you objective?
S: I don’t know whether or not I am a robot.
H: What would you like to do with your life?
S: Let’s talk about movies
H: Have you seen Avatar?
S: More and more all the time, I haven’t seen it. What did you like about Avatar?
H: One thing I liked about Avatar is how these artificial characters really seem alive and artificial characters that seem alive would include you.
S: Thanks for the info.
Sophia’s expressions are very humanlike, but seem strongly rooted in vague annoyance. Her somewhat flat tone seems to hint of sarcasm, which can’t be true. The AI hasn’t advanced so far just yet. Not as far as we know.
At the conclusion of the discussion, the crowd moves forward to take pictures of Sophia. She looks around, with the same expressions she always has. People ask her questions and she gives vaguely sensible answers. Eventually, Sophia is unplugged and all her movement ceases.
My second android experience happens on another sunny day in Austin, Texas. The Palm Door on 6th street has been transformed into the Japan House, a temporary center of culture and technology. After stepping into the dimly-lit building, there is somewhat jarring sound of oddly-mechanical childlike voices. Follow the voices into a small alcove on the right, and you will be greeted by tiny robots with giant, cartoonish, eyes staring at you and awaiting some response. If you linger in silence, they simply chat amongst themselves.
To the left, a gentleman sits at a wooden table. He is well dressed, with shaggy dark hair that brushes just over his bright brown eyes, and rather stern expression. There is a line of people waiting to speak to him. One by one, they sit down across from him at the table, and lean forward to speak into a plant. He looks oddly familiar. He is a robot.
He is familiar because this particular robot is a “”Geminoid, a reference to the Latin word “geminus,” meaning “twin.” This Geminoid is modeled after is creator Hiroshi Ishiguro. Ishiguro recently gave a session talk in the Austin Convention Center at SXSW Interactive entitled Androids and Future Life. In his talk, Ishiguro discusses the state of his robotic innovations and their possible places in human society.
Ishiguro states “In order to create the robot society, we need to have research, studies on humanlike robots.” He goes on to say that “if we want to use robots in society, in daily life, the most important technology is interaction with people. ” Throughout the presentation, he places particular emphasis on observing and interpreting the way that humans interact with his robots. It seems one marker of his success is the human acceptance of his creations.
Human archiving is one theme of these androids. Ishiguro discusses the concept of humans as “national treasures” in Japan. He describes the making of an android designed to look and speak like a particular very traditional “comic storyteller.” The original storyteller is too frail now, to tell the stories like he used to, and so they have “archived” him in the form of an “Android whose appearance is probably younger than him.”
Apparently, the android is more popular than the original form in certain circles. Ishiguro mentions that ” usually the younger people are not interested in traditional comic storytelling.” But this robot has caught their eyes, which Ishiguro explains by saying “The android can have much stronger presence than the original.”
He then segues into questioning the very nature of human presence and tells stories about beautiful robot girls in store windows. According to him, the original condition of mannequins was a waxen hyperrealism and “by using our technology, we can go back to the original idea.”
In Japan, there is a beautiful robot girl that watches visitors in the shop that is her home. She smiles and “always, the people smile back to the very beautiful android,” which he finds quite humorous. He talks about android shopkeepers capable of answering the usual not-so-complicated questions of shoppers. These androids are apparently not as intimidating to some of the shyer shoppers in Japan and are easy to approach. These shopkeepers “give many positive opinions. You are so smart. The clothes are perfectly matching you. You should buy this,” and are apparently quite successful.
Ishiguro delves into the concept of the “uncanny valley.” This is the idea that robots that are almost human, but not quite, are very disturbing, while obviously not human robots are not as bothersome. He talks about how his hyper realistic androids make their way out of the uncanny valley by being so realistic that people accept the humanness of them, and are not so disturbed.
He also discusses the less human-shaped robots. Their voices are human, but their form is only a vague impression of humanity, a neutral appearance. This neutral appearance allows the projection of humanity through imagination and he claims these robots are particular good for the elderly and dementia patients These neutral robots trigger engagement without pressure.
One of the neutral appearance robots is the Hugvie. This is a vaguely human-shaped pillow that a cell phone is inserted into. People can make phone calls and robot generates vibrations that are a reproduction of the human heartbeat. The vibration varies according to the tone of the speaker on the pocketed cell phone. Ishiguro explains that this can reduce cortisol, a stress hormone, because “You have a very strong feeling to hold another person’s body. This is very very strong.”
Throughout the session Ishiguro explains that all attendees must go and visit the popup “Japan House,” for closer interactions with the robots and to have discussions with them. And so I have.
I follow the mechanical childlike voices to find a circle of three CommU robots having a discussion amongst themselves. They are asking each other whether sushi is good, but conclude that electricity is enough. They turn to me with wide eyes waiting for some sort of response. I say “Hello” and they say “It’s great talking to you.”
I move on to greet the Geminoid. It’s strange to see this robot after hearing a session from its inventor and twin, Hiroshi Ishiguro. There is a flower with a microphone in front of the robot that allows it to “hear” the questions of its visitors. To the right of the robot, a man sits at the computer that is plugged into the Geminoid. Occasionally, he chuckles at some response the android gives.
I sit down to speak to the robot. It’s quite disconcerting. The eyes are bright and glassy. The muscle movements are precise and engaging. The voice is humanlike, similar to that of its creator, but it has a mechanical inflection that immediately distinguishes it from that of a human. Some of the responses seem vaguely nonsensical, but such close proximity and humanlike facial expressions lend meaning to the phrases.
The following is part of our discussion:
GEMINOID: Today is the day.
LAURA: Today is the day for what?
G: Ah huh, a crew of angels ceases. Would you like to sing karaoke?
L: Why don’t you start, can you sing something?
G: Ah huh, well who does something?
L: Who does something? That’s a good question.
G: Ah huh, right, so what is your answer?
L: I don’t know, what is your answer?
G: Right, uh huh, I don’t believe so.
L: I don’t believe so either, are you having fun today?
G: Please pick one right, yes, I am, are you having fun these days?
L: Yes I am, meeting such a nice, friendly, robot.
G: Right, well I would like robot for companionship but definitely not humanoid.
L: Well, uh, what’s your favorite robot?
G: Okay, I see Claude Shannon, an American Mathematician, said “I visualize a time when we will be to robots what dogs are to humans, and I’m rooting for the machines.” Can you visualize such a time?
L: Well I’m a little biased, so probably wouldn’t prefer that. As a robot do you advocate the demise of humans?
G: Right, okay right?
L: So a little bit?
G: Demising? Why do you think people are so fascinated with robot takeovers?
L: I don’t know, what do you think?
G: I see, from The Terminator to The Matrix machines, being in charge is never a good thing right?
L: I guess that depends on your perspective, if you’re a robot that might be good?
G: Right, well right. Can you picture droids like the ones in Star Wars in our world?
L: Uh, yes, how about you?
G: Well, well, right. Me?
L: Yes, you.
G: Okay I see.
The robot seems friendly. There is a certain warmth to the creature. Perhaps it is just very well sculpted. There is a group of people gathered around to snap photos. The creature has succeeded in capturing their fascination, and engagement. They strain to catch the Geminoid’s every word. No doubt Ishiguro is pleased and amused his robot’s popularity.