Harshit Agrawal — “While the human remains the central driver of the work, the locus of creativity shifts between human and machine in the process of ML-led art and design.”
Harshit is an artist working with technology. He creates art and builds tools to study how technology can blend with and enhance human creative expression, and in the process create experiences that invite us to reflect upon our relationship with technology. Broadly, he likes working with emerging technologies to poke at creative experiences with them and also actively engages in design and human computer interaction research. He is currently working as a Designer at the Adobe Design Lab and is an alumnus from MIT Media Lab.
What according to you is machine learning led art and design?
In essence, it’s a new form of expression, establishing what I refer to as the ‘human-machine creativity continuum.’ It melds human intentions with machine imaginations, shifting the locus of creativity repeatedly between human and machine over the course of creating a work of art or design. In that sense, it goes somewhat beyond what the inherent notion of a tool is, wherein, it does not simply execute a capability perfectly, but additionally leaves enough room for the machine to create unexpected outcomes, which the human could then incorporate in their overall formulation of the work. Therefore, while the human remains the central driver of the work, the locus of creativity shifts between human and machine in the process.
What role does inspiration play in the process? What kind of inspiration do you seek? Who inspires you the most?
Like in most activities of expression, inspiration plays a huge role for me too. A large narrative of my practice is for it to be a commentary on technology itself- therefore most of the trends of technology- what’s new that researchers are cooking up, how is it affecting people, how are large companies deploying it, in what all ways can it be potentially used, what are conversations around it etc. form a large part of inspiration or material for thought for me. Several artists’ works inspire me too — starting from a lot of the international AI art community today (though it’s a small community as of now- which is a nice aspect in a way as you can interact with most people, it’s nice to have conversations with them and see their works), art studios like Rhizomatiks and artists like Daito Manabe for pushing the boundaries of what art can make you think about and a lot of people who are working actively in the technology ethics space too.
I continue to draw inspirations from the HCI community too, which has been the venue for my academic research over the years. Like I mentioned, I actively follow the discourse on academic research in the ML space too- reading up papers, following conferences etc. Another seemingly unrelated topic that I draw a lot of inspiration and feels from is poetry. I love reading, writing and thinking about poetry- as Emily Dickinson very beautifully puts it- as a way of telling all the truth but telling it slant. I think poetry offers a lot of ways of thinking about art for me, and I’ve ended up trying to incorporate poetry into my works too.
What are the liberties and constraints of computation as a material and process?
I think the liberties are definitely around the large scope of possibilities computation suddenly opens up- almost everything is game in a digital environment! As a process too it’s very interesting to work as a computational artist, especially centered around machine learning technology because of the shifting role of donning a scientist or engineer hat along with the artist hat you don throughout the process. There are stages in the process where you are figuring out the technological details, keeping the ultimate art experience for the audience in mind and some parts where it’s more purely about the tech- getting something to work at all, setting up a server etc!
The time scales of working with ML are particularly interesting too- because you often have to wait at least a few hours for a model to train on your data to get a sense of what it’s throwing back at you. That might work, or you might have to restart training from scratch, therefore I think embracing patience becomes key too! Be it tuning hyper-parameters of network training for a certain aesthetic outcome, to treating the data set collection as a process of art making, while ensuring technical constraints of sizes, number of samples needed etc. are satisfied, I think the crossover between art and technology is quite strong. I find the process very exciting overall- almost like this uncertain journey of figuring stuff out as you go along, remaining open to wandering to new directions and destinations.
What role does the Indian context and narrative play in your work?
As one of the only artists from India working actively with ML technology as a core of their practice, a lot of my work emerges to be around themes that marry Indian visual traditions with ML, be it in the Masked Reality series or Machinic Situatedness or some new works I’m currently doing. This is to think of the broader question of what the cultural underpinning of AI art is. A lot of the AI art that emerges is of Euro-American aesthetic as a result of most AI artists being from that region and most datasets of art available of that kind, and I want to break that by making art that resonates with India more- bright, colorful Buddhist paintings made with AI, or interactive pieces, sculptures and paintings driven from face painting rituals and mask cultures of India. It’s very interesting for me to build this abstract representation of concepts that predate us many hundreds of years, and bring them forth in completely new ways for people to engage with.
Apart from that too, I think Indian mythological creatures in their various body configurations and capabilities are very inspiring for what a non-standard body identity and machine enabled capability (through robotics or in virtual environments) could mean. For example, for the drone artwork (A Flying Pantograph), that was one of the points of inspiration.
How do you think we can inspire more people to explore ML-based tools?
I think the idea would be to get people to see the magic of it- really excite them of the possibilities of what it can do and the questions and themes that you can bring forth with it. Then it’s all downhill from there. There’s obviously always the challenge of how technically sound one needs to be, but there are tools like RunwayML, Teachable Machines etc. and a lot of online learning resources now that are reducing entry barrier and I think once you start dabbling with these, you sort of are likely to get hooked onto them and take it from there. It always works for me to get excited about some new possibilities enough to then Google a lot and figure out how to get it done!
In my capacity, I have started doing many more workshops around this topic too and quite a few of them happen in Bangalore since it’s where I’m based. Would love for people to keep an eye out for that too. Here again, the idea is to inspire, and then give enough basic understanding of concepts and tools to start experimenting oneself.
How has the city of Bengaluru played a role in your career as an artist and a designer?
Bangalore has been fantastic as a city to call home for the last couple of years! It was a very interesting stage in my life as almost a couple of years when I was making a transition of geographies, work and lifestyles at large. I had finished a Masters at the Media Lab, MIT in 2016 and then decided to carry out art residencies for about an year and a half in different parts of the world, immersing myself in new cultures, their studios and museum settings. After that I returned to India towards the end of 2017. By then I wanted to start living in a city for a longer period of time, and not move around every 3–4 months, as I’d done over the residencies.
That’s when I decided to move to Bangalore because a lot of my college friends were here, and because it felt like the city would offer a right mix of technological discussions with cultural influences- and it’s been great! I love that I can walk into most groups of people gatherings in Bangalore, and tell them that I make art with ML, and a lot of people will connect with it in one way or the other. Otherwise, I feel the narrative of technology driven art is quite nascent in India (though it is improving), but in Bangalore I’ve felt people embrace this quite well. I spent a few years here as a kid too, at St Joseph’s and those were great fun- lots of nostalgia for me too from the city. And I am a big fan of South Indian cuisine and filter coffee!!
And finally, to wrap up — How do you spend a perfect Sunday in Bangalore? Any recommendations or must-trys?
There are some spots that have become a regular hanging out place for me- the Bangalore International Center with their wonderful programming being one of them. It’s a great boon that I stay hardly a kilometer away, otherwise you know how traveling in Bangalore can be. Apart from that, a visit to Lal Bagh is always nice, especially if I can get myself to get up in the AM. The combination of Koshys (especially their cold coffee and palak paneer!) and Blossoms is great too!
Find Harshit’s work on:
Instagram and Twitter @harshitrnnh
LinkedIn Harshit-Agrawal
This interview is a part of a series of conversations with designers from Bengaluru called ADI Diaries. You can find more interviews here. Follow ADI Bengaluru on Instagram at adi.bengaluru
