PROMPT: WAR STORIES is a project developed using GPT-3 and various pre-trained text-to-image models [DALL-E 2, stable diffusion], through a series of text prompt inputs chosen by the author.


These pre-trained models consist of quantified traces of reality and collective histories, becoming a substance for algorithms to generate content that recycles the past.

On the borderline between sensitive content and an easy slip into topics of violence, this project explores the depths of the subconscious of these models, excavating the influences of media and online information exchange. These models build the spine of quasi-historical narratives simulated by the algorithms and often obfuscated with prejudice and misinformation, along with the author’s personal bias.

lack of agency, oversaturized media, and perpetual state of war - all of that as part of our collective histories now being reiterated into two imaginary narratives - one more fantasy like heroic context while the other is a reality simulator.  work is proposing a colonised imagination, inviting us to sit on the safe side of this tent--membrane--information shelter and meditate on the generative ideologies of war

Ars Electronica Blog

Barbara Jazbec, Kristina Tica
Speculum Artium, New Media Festival, Trbovlje

Installation view
POWER STRANGERS is an interactive real-time video and audio installation. It consists of a web camera, projector and a water tank filled with reactive slime. It is based on custom-made face detection algorithm. When approaching the installation, the spectator’s face is being traced by a live camera feed. The data received from the camera is being processed and projected inside the tank, only when the spectator turns away from the installation - therefore, they are never able to see the projection of their face.
Lurking behind the spectator, there is an image of them disconnected from the self, yet dependent on it - becoming their ghost. Contemporary ghosts live in the digital data and they generate the power of the automated surveillance systems. Extracting the disembodied - virtual presence from its real life experience, introduces the idea of being haunted by your own ghost - your own digital data. In context of hauntology and retro-futurism, it aims to present current state of commodification towards global surveillance capitalism. Referencing a pop-culture phenomenon that marks the 90’s generations, whose youth seems to be the last one to grasp pre-internet or pre-social media times, the aim of this project is to suggest what comes as a phenomenon of the post-internet world. In this world, the ghosts that exist do not come from a haunted house, they come from our own digital traces. We are all strangers whose ghosts roam around in the world of data, and we generate the power of that world. The more we learn about power that we have to shape this world, the better rangers for the future we are.



real-time generative video installation
computer vision, artificial intelligence, apophenia, pareidolia, cold reading

Ars Electronica - Interface Cult
Man, Machine and Trust, Ars Electronica Blog
Kronen Zeitung


FUTUREFALSEPOSITIVE is a project based on StyleGAN and object recognition algorithms applied to the ritual of Turkish coffee mug reading. A collection of these pictures composed the initial database out of which new images have been generated. 15000 real-life and generated images are then morphing into an animation and used to train the algorithm to recognize objects out of these random shapes created by the coffee stains and generated noise. The algorithm performs this continuous object recognition process in real time—reading the mug—while producing new visual narratives in a loop.
In this process the relation has been established between false positives in computer vision and psychological phenomena of pareidolia and apophenia. The interplay between prediction as a false positive and prophecy as apophenia - the tendency to perceive meaningful connections between seemingly unrelated things - does not only focus on absurdity but on possibilities of creative interpretation when trying to understand the technical processes behind.

*special thanks to Miloš Trifunović and Thomas Hoch



Project supervisor: Uroš Krčadinac

Programming: Miloš Trifunović, Mateja Stevanović, Jacques Laroche
Sound Composer: Marija Šumarac
Supported by: Faculty of Media and Communications, Belgrade
                           Center for Promotion of Science, Belgrade


Festival S.U.T.R.A. // Official Website / September 2019
Remont Gallery / Official Announcement / June 2020

art+science lab - Intelligence IO / September 2020 +official website
Ars Electronica Flickr

The Practice of Art and AI, Ars Electronica, Hatje Cantz, 2021

Uroš Krčadinac - Data Vizantija (essay, SRB)
Uroš Krčadinac - Mentorship



Digital Prayer is using one of the computer vision methods for image generating, with final visual output that resembles an orthodox icon. Through pixel generating using Machine Learning techniques and Generative Adversarial Neural Networks, a relation has been established between the canonical structure of an Orthodox icon and the image artificially created through computer program.
A computer program was created and a database of about 4000 digital reproductions of orthodox christian icons was collected, on the basis of which the computer learns to create a completely new image.
In relation towards geometric and mathematical order of the elements inside of the digital reproduction of the orthodox icon, it is expected that, through statistical operations, the computer will we able to repeat and reproduce the elements of canonical imagery of visual language of the icon, which from the viewers perspective would be perceived as symbolical elements.

When all of the aspects of the virtual environment of the digital image in the age of internet are taken into account, this projectional form of appearance on the surface such as the computer interface, it could be claimed that every phenomenon can be translated onto that projective surface. Icon has had the same purpose during medieval times - as a window towards immaterial, projective, spiritual. The image that is seen is just a visualisation of the code and data, which leads us to the analogy proposed by Boris Groys: “The digital image is a visible copy of the invisible image file, of the invisible data. In this respect the digital image is functioning as a Byzantine icon—as a visible copy of invisible God.
The goal of this artistic project is the demystification of the power of the AI. What used to be the God in feudalistic medieval times, it is AI in today’s neoliberalism. Through direction of algorithmic processes towards imitation of the language of the icon, two visual interfaces oppose to each other - virtual world of digital image, and the traditional handcrafted Orthodox icon.

Intelligence IO - Art + Science Lab
2020, MKM, Belgrade
Photo: Ivan Zupanc / CPN,
Kristina Tica, private archive, 2020


A collection of  text and image content,
fully generated by text prompts
using OpenAI GPT-3 model, VQGAN+CLIP,
allenAI, inferkit, DALL-E mini

Pictorial and textual experimentations on
excavation of the subconsious of a database.
All of the tools for text generating wer

e used as pretrained models.

Considering the intransparency of the databases and models commercialised
for a broad use in research and industry,
the aim of the project is to hack possible interpretations and representations
that might be held within these models.

With many ‘sensitive context’ warnings, these systems are auto-censored
regardless of  the bias that became to complex to be addressed -
the one that has been concieved by the logic of
‘there is no better data like more data’,
where correlation and causality
split the world into two different logics.

The rupture is there, however, one world tries to inject itself into the other
- against technosolutionism  -
some kinds of implemetation of computing processes and automation
cause more  issues than problem solving.

Where is the line between the complexities of  
social, ethical and moral issues
and the optimisation of life?

   ︎  ︎  ︎  ︎