Janaki Vivrekar, the mind behind the Quarantined AI project and a graduate student at UC Berkeley EECS, spoke with Fabiana Jung from Life in Quarantine on August 31, 2020. The following are transcribed excerpts from that conversation. They have been lightly edited for clarity and omit off-the-record comments and asides.
FABIANA JUNG: Hello Janaki, thank you for taking the time to meet with me today. Could you give us a brief introduction of yourself?
JANAKI VIVREKAR: Yes, thank you. My name is Janaki Vivrekar, and I am a graduate student at UC Berkeley EECS at the Hybrid Ecologies Lab studying human-computer interaction and new media. I also recently graduated from Cal where I did my undergrad. The past few months in quarantine have been a great opportunity for me to think about how our lives are changing, and how we can take advantage of this unique context that we are in to produce art, work on some cool projects, and have fun.
FJ: Given current circumstances, it’s admirable how you started a project that reimagines our relationship to the outdoors and indoors. What led you to start @Quarantined AI?
JV: In my worldview, this COVID-19 quarantine period has changed our living contexts in different ways for everyone. So for myself and for many others, the past few months have involved living mostly indoors and being really cautious when we step outside of our living quarters. This led me to question how our relationship with the outdoor world would evolve after quarantining indoors for several months, especially at the beginning of this quarantine period when there was so much uncertainty about how long the quarantine would last since things were changing so quickly. Now that we’re using technology so heavily to stay connected with each other, I was curious about how our machines are “experiencing” the quarantine along with us. This led me to think about how we might, in particular, examine the capacities and limitations of artificial intelligence to creatively parallel this dialogue between our quarantined present and our hopeful future…I only started to think more deeply about this project after I attended a talk at the Berkeley Center for New Media on April 16, 2020, about how advancements in AI and robotics have been fueled by corporations and venture capital, and that these very advancements challenge our assumptions about the distinctions between humans and machines. Christiane Paul, the Chief Curator and Director of the Sheila C. Johnson Design Center talked about “robo-exoticism” which characterizes a range of human responses to AI from fear to fantasy, and this was interesting to me because I was thinking about the potential that AI had to parallel or mirror the human experience. I had this idea and thought, “I should just try this, I’ll learn new skills and try to make something fun out of it.”
FJ: You’ve also recently completed your undergraduate studies at UC Berkeley. Was artificial intelligence something that intrigued you ever since your time in undergrad or was it something that you’ve recently started exploring?
JV: As a Computer Science and Applied Math major, the technical aspect [of AI] has always been interesting to me. I enjoyed the classes I took related to the field, but I realized that a lot of research focuses on making this technology more effective and faster in different ways. For example, there are research groups at UC Berkeley that focus on building more humanoid systems or building AI systems that can do human-like things in different ways. What I’m very interested in right now is the contemporary discussion on how AI can mirror humans and the limitations of that as well. I really enjoy thinking about this theme and I think there’s still a vast opportunity to pursue research on AI through an artistic mindset.
FJ: I find it interesting that you’ve decided to fuse AI and art together since they are generally regarded as completely different fields. What inspired you to combine AI and art together?
JV: The outputs of artificial intelligence systems can be excellent mediums to examine what currently is and what could be, so the present and the imaginary. I first started studying artificial intelligence and machine learning in a computer science/engineering context that had applications to tools and problems in the tech industry. It became evident to me that these systems can be an unexpected source of imagination on topics where we rarely think about applying them. I discovered that there’s even a big AI art community on Twitter and there are a lot of AI artists out there doing interesting and cool things using these systems that are otherwise mostly used to make profit from digital users.
FJ: From the AI art communities that you are a part of, are there any particular AI artists who inspire you?
JV: One of my inspirations is Refik Anadol, an AI artist who uses soundscapes and visual art to build these interesting 3D data sculptures. His work made me think about the potential of taking Quarantined AI beyond the digital image in the future. I got to hear from him at Berkeley in October 2019, and he called his work “data dramatization” — he collects data from nature, like wind currents, and builds these art pieces that showcase data about the universe in different ways.
FJ: Out of the artworks that you’ve posted on Twitter, which one would you say is your personal favorite?
JV: I would say one of my most favorite pieces is Day at the Museum. I really love this one because it originated from an image of a woman sitting at the beach, and putting it through the [Quarantined AI] system, you get something completely different. To me, the final result looks like a woman sitting at a museum in front of a really big painting. I feel like this piece captures a part of my state of mind in quarantine as well as my desire to appreciate and create art given our circumstances. When I see the images outputted by the Quarantined AI system, I feel inspired because I see that there are so many ways to be imaginative and creative and not be stuck in the present but rather use our present context to produce art. I feel like the outputs of AI systems show us capacities of AI to parallel the human experience and also open us up to more possibilities, open us up to imagine more futures and also reimagine the present.
FJ: Along those lines, how do you view your own work?
JV: To me, my work is a learning opportunity. I have been inspired by so many prominent AI artists who are doing fascinating work. Stepping into this domain was completely new to me this summer, so I wanted to challenge myself and let myself be inspired by the community of AI artists I admire. I used this art collection as an opportunity to not only learn new technical skills but also think about the message that I wanted to share and about how we can all add value to the world around us.
FJ: You seem to have quite a developed artistic eye. Do you have an artistic background?
JV: I feel that art has been a part of my life since the very beginning. I’m a musician and I used to draw a lot as a child and I still love to draw and sketch. I enjoy thinking about colors and I really love going to museums. My parents always took me to museums wherever we would go someplace new to learn about the history of that space and the time and era captured in that space. To me, appreciating art is very important but I also feel that art and creating art is always a testament to how we value more than what exists in the world and humans have a desire to add value to our world. And even though there is no definitive answer to what might be of value, it’s important that we keep on searching for it.
FJ: With @Quarantined AI, you’ve transitioned from a spectator at the museum to a curator of a virtual archive. What does it look like behind the scenes of the AI art process?
JV: So with creating art through AI, I learned that there are different stages in the curation process. There’s an initial image source, where I choose pictures that I feel could be relevant to this time. I sourced my original data sets from the MIT Indoor Scene Recognition data set and also from a database of landscape images I found on Taggle. These images were easy to find on the internet. I had a second phase where I sourced data from individuals on Twitter, mostly friends, who had taken pictures of their indoor or outdoor spaces, and I transformed them to see what they would look like using the Quarantined AI system. The system itself is by generative adversarial networks, and in particular, I’m using cyclogan systems to apply a transformation to indoor images to look a bit more like outdoor images. For instance, give it an image of a beach and have it reimagine it to look like a living room. Similarly, after we get image outputs from the system, you can choose which ones you like. The ones I posted on Twitter were just a handful from the many, many images that were generated. I chose select images that evoked some sort of emotion or imagination that I wanted to share.
FJ: What do you hope to gain from this project and what is your hope for others when they see your work?
JV: My hope for this project was to help us reimagine how we might exist in an outdoor context in the post-covid world, if that exists, and if that will be distinct from the current or pre-covid world which were a lot of the thoughts I was having in April when I started this project. One of my goals is to take a participatory angle and source data and images from people in all sorts of living contexts and then train AI systems to portray a more accurate representation of our quarantined lives. The pandemic has disproportionately affected some members of our society in different ways. For example, the homeless population which is growing in the Bay Area and becoming a bigger issue. So then, what opportunities are there for this project to speak to different populations that aren’t experiencing this dichotomy of an outdoor and indoor living space in quarantine?
FJ: What is your advice for someone going into AI or interested in becoming an AI artist?
JV: As someone who’s also a beginner in the field, I would just say “go for it.” It’s easy to look around and be intimidated by this vast world of beautiful art and beauty. One thing that’s important to know is that any art that you make is unique because it’s made by you. There’s nothing that you can’t learn and the more that you can bring your own creativity and mindset to your own work, the more compelling it will be. It will also help others engage more with your own work when they glimpse a part of you because they will be more likely to relate and connect to it. I think this applies to any type of art, music, painting, or AI art. If you’re a student looking to start a new research project, I would recommend getting funding from your project through opportunities at their university. I was able to receive funding from the Koret Undergraduate STEM Research Scholarship. Find mentors, one thing I did was I reached out to people who had experience in the field and also who were open to mentoring it. I built a little slide deck with my idea and what I wanted to accomplish and requested mentorship from them. Sometimes when you have an idea at the beginning, it seems so out of reach but once you verbalize and force yourself to write it down or share it with others and there are concrete steps you can take to accomplish them.
Interested in learning more about Quarantined AI? Check out more of Janaki’s work here.
How has quarantine altered your perspective on the in- and outdoors? How do you feel about the transformations happening to the space you exist in?
FABIANA JUNG – is a Life in Quarantine member. She enjoys hiking, watching movies, and baking.
Center for Spatial and Textual Analysis (CESTA),
4th floor, Wallenberg Hall (bldg. 160)
450 Jane Stanford Way
Stanford, CA 94305
Stanford Mail Code: 2055