Reading Time: 3 minutes
The duo behind CROSSLUCID discuss the particular collective’s latest project Landscapes – a series of 5000 pictures created in collaboration along with generative adversarial networks (GANs) to imagine the shapeshifting life forms of tomorrow
Lopsided skulls. Mercurial skin. Bionic body parts. The ageless entities in Landscapes look, at once, primordial and fantastical, well known and foreign, of nature and artifice. The continuing portrait series is the recent project of the Berlin-based interdisciplinary artist collective CROSSLUCID , created inside collaboration with Generative Adversarial Networks (GANs) and two “data alchemists”, Martino Sarolli and Emanuela Quaranta.
“The outcomes of GANs usually seem uniform because the insight used to train Artificial Intelligence (AI) models are mostly portraits of people who are white, gender-conforming and have conventional Western-looking functions, ” says Sylwana Zybura . Zybura is a founding member of CROSSLUCID , together with Tomas D. Toth . “It was exciting for us that this kind of incredibly large spectrum associated with characters emerged from the nerve organs networks for our new project. ”
The series began in November 2020 after Slanted magazine entrusted CROSSLUCID to create 5000 exclusive covers for an issue delving into the impact of AI on design and the daily lives. The project further speculates on the several underlying structures that notify our identity and how they could emerge in the life types that will populate our potential future world.
Landscapes builds off of the collective’s final photobook, Landscapes Between Eternities , published by Distanz Verlag in 2018, which visualises otherworldly figures and forms flourishing in a future that exists beyond binaries. The guide investigates the many ways all of us construct identities and perceive bodily expressions through pictures of humans melding along with props and shots associated with elusive objects.
CROSSLUCID met Sarolli plus Quaranta in 2019 while working on a project with the Istituto Italiano di Tecnologia in Genoa, Italy. “We experienced many fruitful conversations together about the potentialities of AI, specifically GANs, and provided similar goals in terms of exploring this field, ” recalls Toth. For the new portrait series, CROSSLUCID employed datasets originating from Landscapes Between Eternities . They prepared these through artificial nerve organs networks, which are the processing systems fundamental to heavy learning algorithms, which, since the name suggests, are motivated by models of the human mind. GANs, specifically, are a type of AI model that makes use of two neural networks, which usually compete with each other, to generate output.
“Normally, you need a big dataset, like a minimum of 10, 000 images, to feed GANs so they can look for patterns and find out from them, ” explains Zybura. “Of course, we didn’t have that. We at first had an extremely small dataset of 100 published images from the photobook. So we idea, ‘What would happen if we used test shots from our motion and texture studies since input? ’ That’s when we decided to add around 200 images to the dataset that didn’t make it into the publication or were error shots. From an artistic perspective, it was interesting to show so much of our creative procedure and include these behind-the-scenes images, but , for the GANs to be trained properly, it was also necessary. ”
During the training period, which lasted about five a few months, the two data scientists sent CROSSLUCID a batch of recent outcomes every couple of weeks. The particular artists observed that the preliminary rounds of portraits had been abstract, barely resembling human being forms. Although Zybura and Toth had to adhere to recommendations from the magazine publishers concerning the cover images’ compositions, they will decided early on that it was essential to let the training process occur organically instead of trying to control the visual outcomes by driving it in a specific direction.
“Ultimately, we all didn’t want the final collection of images to reflect our aesthetic preferences, ” states Zybura. “We were keen on showing the mystery and unknowability inherent to the procedure for collaborating with GANs. ”
Uncanny, diverse and showing up as if in perpetual movement, the 5000 images that currently compose Landscapes will remind us that AI is definitely far from neutral. Indeed, a lot of functioning neural networks do not reflect the newest research or acknowledge that learning furthermore happens beyond the brain. “The outcome of these AI types is completely related to their insight, ” says Toth, “To some extent, they can teach all of us about ourselves, our biases and the existing structures that will govern our lives. We hold coming back to the outcomes – almost daily – interacting with all of them, reflecting on them and listening to advice from their various shades and textures. ”
CROSSLUCID veers away from much of the talk that conceptualises AI as being a threat to humanity. Rather, the collective sees these types of emerging technologies as a part of the human ecosystem that can help us picture new possibilities of living plus being in an ever-evolving landscape.