Applied Memetic: Comics Generated by Artificial Intelligence

While speculations about the growing role of machines in artistic production have been a consistent trope in modern and contemporary art debates throughout the 20th century, comics from their early beginnings, have been symbiotically expanding with the development of printing, distribution, communication and media technologies. These industrial processes of completion based on generalized automation, standardization practices and an orchestrated division of labour are so embedded in the ways we understand and consume comics, that have become an essential feature for the conceptualization of artistic practices in the medium. Today, deep neural networks play a transformative role in advancing artificial intelligence across various application domains and some of the most creative bits of contemporary art are happening today at the junctions between different disciplines and technologies. As a consequence, within the “computational creativity” literature, various papers and academic researches have proposed different algorithms and model architectures in the exploration of the creative potential of a machine.
Applied Memetic responds to the need to accelerate the technical affordances in the comics industry. A transdiciplinary team consisting of a comics artist and several computer scientists has the aspiration to produce the first comic narrative entirely generated by Artificial Intelligence. The project represents a considerable technical and artistic challenge as it explores a set of unconventional operations that don’t account for the production of comic books: web-scraping, image classification, computer vision algorithms, language modeling, indexation, database building and cloud computation. Furthermore, the resources from Machine Learning are steered toward the synthetic production of everything related to original comics art: the artwork, the character designs, the dialogues, the narrative evolution, and the page layout will be entirely generated using the most up-to-date algorithmic architectures and models in Deep Neural Networks such as GAN, GPT-2 and transformers. More than a technical challenge, this is an opportunity to explore unconventional processes by weakening the aesthetics predispositions and received wisdom that are reproduced through specific (human) evolutionary interpretations of artistic production. Interested in harnessing the machinic understanding of comics through recurrent patterns, probability distributions and outliers in comics language that have been lurking in the reader’s pre-attentive reader’s cognition and that we haven’t been able to articulate in words, Applied Memetic embraces the machinic volition in the production of an art object in order to unfold a non-human understanding of the comics medium. During our visually-rich presentation, we would walk the reader through a conceptual, historical and technical understanding of the project in our effort to produce a knowledge-rich, experimental transdisciplinary project that pushes the boundaries of the comics medium and aspires to produce the first comic entirely generated by deep learning.