‘Code is Art’: A Chat with Artist Sergio Albiac
Explorations of AI Art — Episode 20
[This interview has been previously published on Cueva Gallery’s website on June 22, 2020]
“I consider code, the algorithm itself, a work of art.” Sergio Albiac
Sergio Albiac is an artist with a computer science and art background, based in Barcelona, Spain, whose work is at the intersection of deep learning, generative computer code and traditional media. Not being constrained by a single medium or style, but interested in exploring the realm of possibilities that technology offers to create art, he investigates how traditional and new media can blend together and bring to life something of profound value. Thus were born paintings, video art pieces, interactive installations running on the cloud, giclée prints and generative portraits.
Fascinated by the dialogues between control vs. randomness, and human vs. machine, Albiac’s work investigates also the concepts of identity, creativity, emotions, memories, and chance, and gives space to the tensions that arise when the reality we build in our heads confronts the external world. Many are the works created around these ideas.
In The generative identity of Walter Vanhaerents, for example, a project and interactive art installation in partnership with Lexus, the artist transforms the car into a work of art. Each time the driver starts a journey, a new artwork is created using real time data from the car sensors. The output becomes a sum, and interpretation, of the artist’s hand brushwork and generative aesthetics affected by the driving style and the duration of the drive. The project — which explores human identity by mixing choice and chance with randomness and control — has been a challenge and has pushed Albiac to further develop a flexible process of generative artistic creation that transforms meaning into artworks.
Another work that explores human identity is I am. Here Albiac concerns himself with the ways in which code can be used as a medium to represent human identity, to tell stories about people. The cloud-based installation, presented during the Internet Age Media weekend (April 8th and 9th, 2016), combines the artist’s generative code with data, cloud computing and real time machine learning services. People participating in the installation were recorded. Then the voice was transcribed into text and the transcription enriched with passages borrowed from literature or philosophy in a free association based on semantic analysis. The result is a series of generative collage portraits which offer new insights and become a portal to human identity.
Through his practice, the artist investigates also the concept of creativity and artistic creation, comparing traditional and new media (and states of mind in the process). For instance, in Generative sketching , Albiac talks about how a painting is a struggle between control and randomness, where the paint’s behavior can not be fully mastered, but only affected by the technique chosen by the painter. The traditional sketching process, useful for inspiration and to test color and composition, starts with a vague idea that later takes on a clearer form as the work progresses. Once this idea is refined, it is put on a traditional medium giving life to a work of art. With neural sketching, instead, an idea in the artist’s head is transformed into computer code, permitting to explore alternatives in a very different way compared with tradition. Writing the code for generating images to use as painting input, according to Albiac, could unleash the (infinite) potential of an artist.
The idea of infinite artworks, not curated by the human artist, but truly novel and created by generative assistants, has occurred for the first time in Stardust. Here Albiac has hypothesized the use of technology to outsource the creation of art. Like he explains: “Modelling artistic decisions into software would provide a generative assistant that could even survive an artist in the creation of meaningful works of visual art.”
The inner and artistic world of this artist is complex and varied. For instance, in Your World is my Landscapes the artists reflects on the connection between words and images, transforming the secrecy of life and private communication (i.e., the words contained in private emails) into almost evanescence lands. Other times, like in Emotional Fields, there is no narrative but pure contemplation of colour.
With us, the artist has discussed his work, perspective and ideas about AI and art. He has also shared unpublished experimental images generated using custom GAN systems that create images from latent space. As Albiac specified to me, most of the time he uses his own datasets which can guarantee a higher level of artistic control. His interest is in the generation of images that look unreal, deeply inspired by representational painting and tradition and where emotion opposes the reason.
Beth: I quote you: “Life is finite but creativity is not.” What is creativity for you?
S.A.: It is a meaningful search in the space of possibilities. It is a rebellion against the linearity of human existence.
Beth: On your website we read that “Code is art” and “algorithm is the artwork”: can you expand this concept?
S.A.: “Code is Art” refers both to computer programming as an artistic medium and the fact that I consider code, the algorithm itself, a work of art: the expression of an idea that has the potential to manifest itself through a vast number of specific instances. Let’s take a traditional art form like painting. A painter may want to express a concept, an idea, a story, a statement and so she creates a specific painting. Maybe she will create a new specific painting around the same concept or idea and so on. But expressing an artistic idea or story through code is the closest attempt to model the abstract thought itself. A thought that with the help of an actor (a computing device), is revealed in the form of one, ten or infinite sensory experiences. But even if the algorithm is never executed, the algorithm formally expresses an artistic intention: it is the artwork. I am particularly interested in those algorithms that embody some degree of unpredictability or randomness. I force myself to give up control on the specific, let’s say, image or sound that the algorithm will produce. It is both a way to make an statement around the human illusion of control in a complex and unpredictable existence. It is the expression of potential, human will, creative impulse and, at the same time, surrender, acceptance and humbleness. It is life force and it is death.
Beth: Much of your work revolves around the concepts of identity, freedom, control, possibility, randomness. Emotions, personal experience and memories are at the center of your artistic research. Do you think these concepts have, in any way, evolved in parallel to your technical and artistic maturation?
S.A.: The medium evolves much faster than the themes. I am still interested in pretty much the same themes I was interested when I started my artistic practice and even before: the blessing and the drama of being a thinking animal, the fabrication of meaning, the distortions of thought and memory, the realization of identity. My artistic journey is an ongoing and never ending attempt to explore these questions through the most contemporary medium today: software. I am not sure that “artistic maturation” is an objective in itself in my case. My practice is experimental by nature. The coherence of my practice resides in the process and the ideas behind it, not in visual language or elements of style. This is sometimes challenging as it might be interpreted as a lack of artistic direction but I’m OK with this as long as this approximation to artistic practice preserves much of the uncompromising freedom that is the foundation of my motivation to do what I do.
Beth: How has the shift from Processing to GANs worked for you?
S.A.: Although I started using Processing in my first artworks, nowadays I use just “code” through multiple programming languages, developing my art with the help of frameworks, libraries or pure code from scratch. Sometimes is practical to use abstractions like Processing or openFrameworks and sometimes you need to build your work directly on the raw abstractions the programming language offers. These are just technical details of the practice, similar, in my view, to using off the shelve tubes of oil paint vs. grinding rocks to obtain pigments and creating your own painting materials. It is a matter of convenience vs. control. But the important thing is the outcome and the ideas behind it.
Regarding a shift from generative code to GANs, in my case, there is no such a shift. I still use generative code and I am incorporating GANs and, in general, deep learning techniques into my practice. More than a shift, it is an expansion of my toolbox. I see more potential in an “integration” of generative code with deep learning and this is one of the strategies I am exploring, along with the more direct use of standalone GANs and dataset design and creation.
Beth: In Latent Empathy you explore the concept of empathy through AI. Is Artificial Intelligence growing as an independent subject of study for you?
S.A.: I am interested in AI as a disruptive technology in relation to the individual and society. And I am interested in AI as artistic metaphor for consciousness, will and instinct. Finally, I am interested in AI as an instrumental technology to model complex generative patterns of outputs that might be useful in my work.
Beth: What is the most exciting project you have done?
S.A.: While you are at it, the project you work on it is “the most exciting project you have -nearly- done”. Any artistic project should be exciting to deserve being considered, otherwise I just don’t work on it. In retrospective, a couple of projects are special though. One is Videorative Portrait of Randall Okita, because it is my closest attempt to express my view on human identity, emotion and memory, and there is a lot of potential in the concept and the technique I developed. The other is Moonracer series, because it is a project still running today and it has produced to date tens of thousands of individual “instances” (or “artworks” resulting from the base algorithm). And the most recent excitement is reserved to a couple projects I am currently working on, using deep learning.
Beth: How would you like to see AI and art progress together in the future?
S.A.: In pure AI research, I’m looking forward to see advances in the integration of machine learning approaches with symbolic reasoning. On a more meta level, I hope we, as citizens, will find a way to press governments and companies to force the deployment of a fair AI and have a voice in the potential tensions that such a disruptive tech will bring to society.
Sooner or later, art built around AI (conceptually or just as an artistic tool) will be considered part of the contemporary art conversation, beyond the current status of PR stunt, novelty or mere curiosity.
Beth: A question nobody asked and you would like to answer?
S.A.: I like the idea of a non-realized question that has not been answered. Let’s just observe this potential without instantiating it…
To follow Sergio Albiac:
Twitter: @sergioalbiac
Instagram: @sergioalbiac
Website: https://www.sergioalbiac.com
References and Resources
https://www.fastcompany.com/90300761/can-you-tell-the-difference-between-rembrandt-and-an-algorithm
https://www.designboom.com/art/sergio-albiac-dutch-oil-painting-learnt-nothing-02-01-19/
About the author: Beth Jochim is the Creative AI Lead at Libre AI, and Director and Co-Founder at Cueva Gallery. She works at the intersection of technology and arts. She is actively involved in different activities that aim to democratize the field of Artificial Intelligence and Machine Learning, bringing the benefits of AI/ML to a larger audience. Connect whit Beth in LinkedIn or Twitter.
