@myfriendsylvia and Aging Virtually

Ziv Schneider reflects on the short life of Sylvia, a virtual Instagram influencer she created and aged over the course of five months. The project capped off with a virtual funeral for Sylvia, as she was fondly remembered by her creators and followers on Instagram. The experiment sheds light on prevalent ageism in digital space and offers an alternative to the mainstream design of virtual humans.

The illustrations include snippets of Sylvia’s direct messaging with a 13-year-old follower.

This past November, Sylvia Novack was put to rest and more than 60 people attended a virtual memorial service for her. The setup was simple but quaint. A cluster of white roses and a framed photo of Sylvia adorned a digital “room” bathed in candlelight. Next to the photo stood a few orchids and a book entitled “What I Learned About Life,” a work she hadn’t yet published before her death. In the middle of the room, a slideshow was projected in  a loop to give guests a glimpse of Sylvia’s life. She was a designer, an author, a fashionista and a muse. She was *obsessed* with coffee and she was a devout yogi. And more importantly, Sylvia did not exist. 

Sylvia was an experiment that I created, a computer generated character that was posted to Instagram from July to November 2020. And, in the course of those five months, she aged from 30 to 80. 

A virtual human is, fundamentally, a computer program. In the far future, it may be something else, but for the foreseeable future, a virtual human is simply code and data which has been designed, and may be evolving, to give the illusion of a human being.¹  

Different forms of virtual humans are becoming rather common. Some use cases include beings with conversational abilities, such as chatbots and virtual companions. For example, the app Replika allows the user to chat and develop a relationship with a customizable virtual human friend. When customizing an AI friend, users can choose their race and gender(non-binary is also an  option), but one thing that all the options have in common is that they’re all young. 

Virtual Influencers(VI) are on the other side of the digital intelligence spectrum. They have no intelligence built into them. If you follow virtual influencers on Instagram, it can be quite confusing finding someone like @lilmiquela in your feed. And depending on your level of media literacy, it can take some time to understand what you’re looking at, and decipher how the image was created. Virtual influencers are rendered characters that live on social media right now and none of them are automated. Some of these accounts use terms like “robot” and “AI” to self describe, which creates a degree of confusion. As a platform for competitive daily performances of our curated personas, where the lines are blurred between private and public, personal and professional, real and fake, Instagram has become the perfect stage for virtual influencers to shine and that community has contributed generously to the platform’s phenomenal success.

The trend for businesses to partner with lifelike CG characters to model their products has been around for roughly four years, and some suggest that the pandemic has accelerated their popularity, as in-person fashion productions aren’t feasible. Without any physical restrictions, these virtual talents can be cast and make appearances anywhere in the world.

Christopher Travers of VirtualHumans.org, an organization that documents virtual humans, points out the benefits of working with VIs:

They are cheaper to work with than humans in the long term, are 100% controllable, can appear in many places at once, and, most importantly, they never age or die.²  

But a number of issues have arisen with virtual influencers, from the lack of credibility of a virtual model advertising beauty products such as make-up made for human skin, to matters of identity and representation, including the virtual black supermodel Shudu.gram that is created and owned by a white man, down to the notion that VIs promote artifice, not  beauty, and disrupt the truth.

When it comes to age, the choice, unfortunately, always seems obvious – make them young, attractive and immortal. After all, aging is not a desirable feature in the American society, which venerates youth and in denial about death. 

The average lifespan of a human being is 79 years. Humans are living longer than ever before, but mass media’s reporting on aging and the aged, is overtaken by the premise that youth is the part of life to be most represented and celebrated. While 46% of the US adult population is over 50, an analysis of online images found that only 15% include people in this age group.³ This exclusion of aging populations is even more prominent when looking at the design specs for virtual humans. Social media platforms like Instagram play a large role in perpetuating body stereotypes and the competitive celebration of youth. 

As we build more virtual humans and see more of them in the world, it’s worth examining what we choose to portray as human by challenging the design concept that initiates the making of the virtual human. What is this person’s gender? What is their ethnicity? And how old is this person? So much work goes into making these beings appear convincingly real. So why is aging’s effect on the human body never considered an essential phase of being human. What could be more real than some wrinkles and other signs pointing to the passing of time? 

After working on Sylvia for a mere five months, I discovered that operating a VI is quite labor intensive. Between the posing, garment fitting, styling, environment design and rendering, one image can take anywhere from one to three days to create, depending on the size and skill level of the team. 

By understanding the amount of labor involved, I became aware enough to appreciate a visual’s well-composed lie. This time spent with the VIs also provided some insights into why virtual beings basically haven’t changed over time or varied body shapes. Any changes made to the virtual body creates a series of tasks and makes it harder for artists to use off-the-shelf digital assets, such as clothes that are designed for presets of body shapes. While excluding aging from the design process seems an obvious choice given the very practical reasons of budget and time, it is still the ruling  every time and designers can easily decide otherwise. One of Sylvia’s posts prompted a comment from a user saying “ROBOTS DON’T AGE”, to which I say – robots will be what you tell them to be; robots are designed by humans. Behind every virtual human is at least one biological human and often a whole team, hunched in front of their computers, possibly vitamin D deficient and overworked. 

Sylvia’s name came from the acronym of the internal project name in its initial development stages – a Silver Virtual Influencer. For me, the term Silver is not only a reference to the elderly whose hair follicles have lost pigment cells. Silver is a gradient, a transition, a non binary. In the digital world she lived in, Sylvia was a glitch, as Legacy Russel writes in Glitch Feminism: A Manifesto:

“Glitch is about claiming our right to complexity, to range, within and beyond the proverbial margins . . . [G]litch is celebrated as a vehicle of refusal, a strategy of nonperformance.”⁴

When designing Sylvia, it was important for me to portray her as authentically as possible based on my own lived experience, while being cognizant of the absurdity of trying to be authentic when posting as an aging robot. Some aspects of her identity were based on my own, like her Eastern European heritage and graphic design background. Since I was the one posting, commenting and messaging as her, I was Sylvia and she was me. And then she died. 

The decision to age Sylvia as rapidly as I did led to her inevitable death at the end of the project. With the announcement of her passing, after five  months of living online, I was taken with waves of grief. The performance of aging and dying was almost a form of therapy for me, revealing the secret of virtual beings as a way of connecting to our deepest fears, and to the even more vulnerable regions of the human experience. 

Before starting this project, I was expecting to see the mostly biased and agist responses from users as Sylvia aged. But to my surprise, Sylvia had teenage followers who loved her for who she was. Some grew attached to her, connected to her story and were sad when she passed. Some people saw her as a role model and sent her sweet and seemingly heartfelt DMs. Some creators of other virtual influencers stated that she has inspired their work moving forward. 

This experiment was about the passing of time, and I think it’s only fitting to allow more time to reveal Sylvia’s influence on the digital world in which she existed. With Sylvia, I proposed an alternative to the way virtual beings, and virtual influencers in particular, are designed. I wanted to offer criticism of this newly emerging practice and point out the ageism that is inherent to the way people design their fantasy humans. By immersing myself into the operation of a virtual human on Instagram, I learned that technology might be one culprit of the homogeneity in the characters currently being made. The power and potential of online performance-enhanced computer generated media presents many concerns but also provides opportunity for making real, human connections.  

Footnotes

1.  Virtual Humans, Today and Tomorrow – David Burden, Maggi Savin-Badent
2.  Virtual Influencers Make Real Money While Covid Locks Down Human Stars 
3.  AARP, Media Image Landscape -Age Representation in Online Images
4. Glitch Feminism, a manifesto, by Legacy Russel, 2020

Ziv Schneider (ITP 2015) is an artist and designer based in Queens, New York. Her work explores the affordances of emerging technologies for storytelling. She is currently a creative technologist at The Brown Institute for Media Innovation at Columbia University Graduate School of Journalism. 

Sylvia was supported by The Brown Institute for Media Innovation. It was also part of the IDFA DocLab R&D program 2020 in collaboration with the MIT Open Documentary Lab.

Styling and Art Direction: Odie Senesh, Character Artist: Halime Maloof, Celebration Director: Bethany Tabor, Natural Language Processing: Alex Calderwood, Digital Installation Design & Development: Tong Wu, Music: Philippe Lambert, Aging Researcher: Alexa Fleet