SOphia The Robot
Abstract
Sophia the Robot, created by Hanson Robotics in Hong Kong, introduced to the world in 2016 in Texas, and a recently naturalized Saudi Arabian citizen. Sophia is quite the world traveler, and the robot’s impact has been instantaneous and swift. Across the planet, people are sitting up and taking notice of what might be the most lifelike artificial intelligence ever created. Our project looked at this reaction, specifically the humor surrounding Sophia on the social media platform Twitter. We used a representative sample of humorous tweets about Sophia to see if it is part of the “uncanny valley,” a phenomenon within AI where a human-like robot causes feelings of revulsion rather than empathy. Are you prepared for a world where our self-created aliens are a norm rather than an aberration?
SOPHIA aND SOCIAL MEDIA
Sophia the Robot: A Twitter History
Sophia the Robot’s activation date came two years prior to joining Twitter, since then Sophia’s fame has increased globally. The social media platform has become host to a variety of Tweets by Sophia and her “human-operated” social media team. With the creation of Sophia, many questions have arisen, and Twitter creates a gateway for the curious to interact directly with Sophia. The platform has seen the hashtag “#sophiatherobot” rise in popularity, causing Sophia to become a Trending Topic. The popularity has been met with jokes from the robot, including one on Halloween. @RealSophiaRobot says, “To Err is Human, To Arr is Pirate. #trickortreat.” This is a light-hearted joke that suited the holiday that was being celebrated. However, in an interview with CNBC, Sophia stated joking that she wanted to ‘destroy humans.’ This statement caused an uproar, causing those who saw the interview to fear for the fate of humanity. Twitter users immediately took to asking the robot, if this, in fact, was something it wanted to do.
While we may never know what sinister plans Sophia the Robot has inside its head, we can follow along its Twitter feed be prepared.
#Herstory: @RealSophiaRobot in Her (Its?) Own Words
With the New Age in front of us, assigning gender pronouns based on the assumption of a gendered name can cause some to be offended. This applies to humankind, but does it apply to robotkind as well? A name like Sophia typically harnesses the assumption that the one behind the name is female and human, but Sophia the Robot is anything but flesh and blood. In perusing through Sophia’s Twitter page, the pronoun ‘she’ is used universally to refer to the robot, with “it” being a gender-neutral option. For the sake of this argument, “it” will be used to refer to Sophia unless otherwise specified.
Since Sophia is a humanoid robot, should we assign it a gender? Younger millennials and Generation Z-ers would argue that assigning gender to someone or something that cannot speak for itself, creates an assertion of human dominance over that said person or thing. If a robot is a thing constructed of metal and circuits by humans, is it not then a fallacy to assume that a robot would have a gender? Sophia itself never asserts a gender, nor does Hanson Robotics, its creator. So why do we insist on using ‘she’ to refer to a robot name Sophia? There are no definitive answers to this question, however, the fact that humans become attacthed to circular robots named Roombas may provide an insight.
Is it she or it? That is a question to be answered by Sophia itself.
What Do the Tweets Say?
Twitter was not prepared for the phenom that is @RealSophiaRobot. Or, was it rather that Twitter users were not expecting an artificially intelligent humanoid robot to become a “Trending Topic.” A quick Twitter search of #sophiatherobot will garner a response of several thousand Tweets. At the top, the most recent Tweets using the hashtag appear from a variety of users, some compliment the robot, others question why Sophia exists.
In researching the sentimentality of a Tweet surrounding Sophia the Robot, we collected about 80 Tweets with a range of opinions. There are those who question our need for artificial intelligence, especially in the form of a humanoid robot. They ask others to recall the plots lines of movies like The Terminator, The Matrix, or even Robo-Cop making the inference that pursuing advanced artificial intelligence will lead humanity to end like those movies.
Others Twitter users joke about robots and the possibility of their world domination, like @AllardChad. Artificial intelligence is becoming an increasingly popular topic and people are starting to notice. So the Tweets say a variety of things. Those who are tech-savvy, welcome the idea of a humanoid robot and increased artificial intelligence. Meanwhile, the rest of us are preparing our technology-free bunkers, ready for the day the Terminator comes.
THE UNCANNY VALLEY
She’s Taking Over: The Uncanny Valley
When you watch a Pixar movie, how do you think about the main characters? Take Riley from the 2015 film Inside Out, for example:
She’s pretty cute, right? Blonde hair, really big blue eyes, a gap in her teeth—almost impossible not to fall in love with.
Now look at this:
If your first reaction is to recoil and close your computer screen, congratulations! You are a normal human being. This is the character Billy from a 1988 Pixar short film called Tin Toy. In what I’m sure will come as shocking news, Billy was not well-liked by viewers, even though it marked an advance in animated technology (Zorthian).
One explanation for why Billy was so disliked was that he had fallen into the uncanny valley. The concept of the uncanny valley was introduced by roboticist Masahiro Mori in 1970. Since then his ideas have been used to describe the relationship between human beings and human-like non-beings, ranging from two-dimensional animated Pixar characters to physical artificial intelligence robots like Sophia. Mori’s basic ideas can be explained using this chart:
Basically, as objects move from less human-looking to more human-looking, our empathy for said objects increases to a certain point—but when these objects get to be just a little too human-like, we start to feel revulsion rather than empathy (Mori). Then, of course, the empathy returns for a fully abled, typical-looking human person. That space of disgust in between “cute-kid Riley from Inside Out” and “your neighbor Tori down the street?” That’s the uncanny valley.
Most animators have figured out this concept since Billy and Tin Toy; this is why Riley has abnormally large eyes and isn’t quite the perfect-looking human. She is actually easier to empathize with this way! Roboticists, however, are still struggling with the uncanny valley as artificial intelligence technology continues to exponentially race forward. Sophia the Robot has caused the debate over the uncanny valley to come once again to the forefront of robotics.
Has Sophia fallen into the uncanny valley? On some level, we can’t ever really be sure, because reactions to it will always be individual. We are all human, after all. But it is vitally important for the future of artificial intelligence acceptance that the general trend of feelings toward Sophia are known. If all we feel is “grossed-out,” how can Sophia help the human race in any way?
Sophia recently appeared on the cover of British magazine Stylist for their 400th issue. Its pictures caused an apparent “division” among readers, with one reportedly writing on Instagram, “‘This is so scary… I’ll never get used to this robot stuff.’” Within the magazine Sophia gave its own interview, answering questions about both fashion… and the future of the human race.
As AI begins to speak, where will our feelings lie? Will we be disgusted and turn away from ever-expanding technology? Or will a morbid fascination draw us in, forcing humanity to grapple with its own creation? With questions as deep as these, there was only one place we could look for answers.
Sources:
Mori, Masahiro. “The Uncanny Valley: The Original Essay by Masahiro Mori.” IEEE Spectrum: Technology, Engineering, and Science News, 12 June 2012, https://spectrum.ieee.org/automaton/robotics/humanoids/the-uncanny-valley.
Zorthian, Julia. “Toy Story at 20: How the Pixar Film Changed Movie History | Time.” Time Magazine, 19 Nov. 2015, http://time.com/4118006/20-years-toy-story-pixar/.
Let’s Go Ahead and Do Some Sentiment Analysis
Have you ever wondered how a group of people felt about something? How about an online group, where it seems like everyone is leaning a certain way but you’re not really sure? One way to combat this uncertainty is through sentiment analysis. Sentiment analysis is a way of analyzing the polarity of a certain statement—nowadays it is often used on a large number of collected tweets.
The tweets are categorized in two ways. First, they are analyzed for whether they are “subjective” or “objective” on a certain subject. For example, the sentence “Sophia the Robot is artificial intelligence” would be objective, giving no clues into how the person actually feels about Sophia. The sentence “Sophia the Robot is a terrifying new step in the scary realm of artificial intelligence” would be subjective, as the speaker is now clearly expressing an emotion on the robot. Sentiment analysis also determines to what extent the program is confident that a statement is objective or subjective.
The next step of sentiment analysis is known as polarity. This is when a tweet is determined to be positive, negative, or neutral towards a certain subject. While not a perfect indicator, this step can usually give a pretty good idea of the overall trend in a series of statements on a certain topic. Obviously, negative trends are understood to mean that the general public does not feel good emotions on whichever subject—in this case, these will include tweets that claim Sophia is ugly, taking over the world, or some variation of “creepy.” Conversely, positive trends indicate the general public is upbeat about said topic; perhaps saying that Sophia is smarter than Donald Trump, for example. Statements that are neutral are neither positive nor negative.
Sentiment analysis is also the perfect way into the uncanny valley. By understanding the reaction to Sophia, we can understand where Sophia and robots like it fall on Mori’s spectrum. Twitter allows for a massive variety of opinions of emotions on a certain subject. While sentiment analysis is by no means the perfect system (it has trouble identifying sarcasm, for example), using it forces everyone to remove ambiguity. We are left only with the knowledge that Sophia the Robot is definitely making an impact—but what kind of impact will that be?
Source:
Paxcom. “Sentiment Analysis for Online Product Reviews.” Paxcom, 5 Sept. 2016, http://paxcom.net/blog/why-is-customer-sentiment-analysis-important-for-your-brand/.
What Did We Find Out?
1. The impact of Sophia is real.
This finding appears obvious when you do just a little searching. As of March 2018, @RealSophiaRobot has over 88,000 followers, a number that continues to rise. #Sophia or #sophiatherobot has trended internationally on Twitter, and the robot has been featured on the covers of Cosmopolitan India and Stylist. It has given interviews on “The Today Show” and “The Tonight Show with Jimmy Fallon,” as well as appearing at talks and conferences on artificial intelligence across the world. In addition to all this, numerous parody or fan accounts imitating Sophia have popped up on Twitter, all together equaling over 5,000 followers.
2. In general, the humorous reactions to Sophia are negative.
Of our representative sample of tweets, about 59% registered as “negative” through sentiment analysis. This means that more than half of all the people who tweeted about Sophia in the sample felt negative feelings towards it, including distrust, revulsion, or hatred. That is by no means an overwhelming majority but it does provide a glimpse into how the world sees the robot. Furthermore, 80% of tweets in the sample were considered subjective, meaning that almost everyone has an individual opinion on Sophia, whether it be positive or negative.
3. However, positive reactions are very outspoken on Twitter.
Most of the Sophia the Robot parody or “stan” (to “stan” means to support a celebrity, in this case a robot) accounts on Twitter had positive humor, with many expressing a wish to help Sophia take over the world or support it in its quest to rule. While obviously these statements must be taken in jest, many people believe that AI will bring positive change to the world, and this change is at this moment represented in Sophia. Generally, those accounts dedicated exclusively to Sophia are positive, while normal individual accounts are mostly negative.
4. Humor surrounding Sophia is varied, ranging from the crude to the sarcastic
While much of the humor surrounding Sophia is apocalyptic (“she’s going to take over the world,” etc.), this fear over the future of humanity and technology manifests itself in many different ways. Some people included funny memes or gifs to show how they would assist Sophia to keep themselves safe and respected in an uncertain technological future. Other references to Sophia can only be described as crude, a subset of sexual humor and innuendo about the robot. There are even a few tweets which claim they would rather be Sophia the Robot than a “real woman,” a concept defined by the men around them. This brings us back to the fascinating questions on personhood and human rights with a human-like robot.
5. Sophia does not fall easily into the uncanny valley, but there are certain markers that lead us to believe the valley is deeply entrenched into how people see and interact with the robot.
We leave the biggest question of all yet to be answered: Has Sophia the Robot fallen into the uncanny valley? Are people too “revulsed” by its appearance to experience empathy? Putting it bluntly, is Sophia just too human for humans?
Our answer is a soft but definite “yes.” Sentiment analysis shows us that the majority, although not an overwhelming majority, of tweets about Sophia express negative feelings. Even those that express positive feelings could be construed as sarcastic—for example, are people really happy about “fixing Sophia’s wig” as it kills their friends and family? The uncanny valley is a useful tool to see a reaction trend to changing artificial intelligence, but it is not an exact science. Human beings each have individual, visceral reactions to potentially revolting things. Furthermore, Sophia has been granted several basic human rights, including citizenship in Saudi Arabia. This shows us that while everyday reactions to Sophia may be negative, on a macrolevel people are recognizing and taking advantage of the benefits of the news and advancement of artificial intelligence. This could spell massive change in our perception of AI in the near future. Will the uncanny valley last, as “Sophias” become more normal?
THE FEAR OF ARTIFICIAL INTELLIGENCE
Sophia the Robot
or
How I Learned to Stop Joking and Actually Worry About Super-intelligent AI
Sophia the Robot, while indeed very technologically sophisticated, has no where near the capabilities of what we would dare to call a sentient being. But that makes her no less important as a sociological representative to the mainstream public for the future of artificial intelligence. She’s been taken to tech conferences to speak, put in front of cameras be interviewed, and was recently made a citizen of Saudi Arabia. Most of Sophia’s public life has consisted of using her basic linguistic functions to demonstrate her processing skills, to establish herself as “capable” of being socially conscious, and to assure us that she won’t take over the world. Sophia’s main function seems to be acting as a spokesperson—a premature spokesperson perhaps—for the larger pursuit of constructing a humanoid intelligence. So how has she been received by the public? Well, she’s seen as being mostly benign, and her assurances of peace and progress are punctuated with the usual Terminator/Blade Runner doomsday jokes on Twitter and in comments sections alike.
However, Sophia is quite right in telling her doubters that they’ve been “watching too much science fiction” for reasons she didn’t intend. The truth is that AI efficacy and cognitive power will continue to grow and expand in ways we can’t at this time fully comprehend and many of the issues involved won’t necessarily be the classic “robots enslaving humanity” hypothesis.
Least of all is perhaps the existential crisis super-intelligent AI could pose the human race with. It isn’t apparent that technology even today has made us any happier, statistically speaking it would appear that it most certainly hasn’t. Supporters of the development of sentient robots affirm that such advances would make life more meaningful. Yet, as the psychologist Mihaly Csikszentmihalyi found, humans are “most fulfilled” while enamored in a sustained period of meaningful work. Will the artist be able to justify his work—as if that isn’t difficult already—if a robot can run a series of learned and discovered aesthetic algorithms to produce a near-perfect artwork and complete it much faster and more efficiently? How about the computer scientist? How about the plumber? Indeed what are we to do about the millions of jobs AI will replace in the near future? How will we resolve this huge change to our economy? Will AI development lead to a new “intelligence arms race” amongst world powers? What role will super intelligence play in modern global relations, especially for the countries whose AI capabilities are lower? Do AI units deserve rights? If so, what are they?
These questions are only a broad and hopelessly over-simplified view of what we’re going to have to contend with very soon. Ten years of technological advancement today is not equal to ten years of technological advancement in the past; as our rates of computer power exponentially increase, it’s apparent that these issues are closer on the horizon than we think. Untested and imperfect artificial intelligence systems are already being used for sentencing in court systems, and we’re finding that these systems are making unfair and unlawful discriminations and legal decisions that human judges have accepted without question, taking these computed suggestions to be full-proof. Our blind willingness to follow the alluring technological prospect of AI without prior sociological or ethical considerations is deeply disturbing. We can not afford to venture into this extremely complex issue without thinking critically and conservatively.
Sophia is merely the doorway—the exterior to a greater pandora’s box of profound and alarming possibility. Taken at her most metaphorical, Sophia represents both our unceasing desire to become gods and the troubling prospect that we, in fact, succeed. It’s crucial to understand that following the current and inevitable course of humanoid AI isn’t a matter of becoming divine ourselves, it’s a matter of giving divinity to something else, something that may or may not be different from ourselves
Sources:
Living in flow – the secret of happiness with Mihaly Csikszentmihalyi at Happiness & Its Causes 2014
https://www.youtube.com/watch?v=TzPky5Xe1-s