Generative art is on the rise. Algorithms that turn text prompts into images, such as DALL-E and Stable Diffusion, appear to be viable creative tools. And they feed many debates on their artistic legitimacy and their potential to deprive us of jobs.
The sudden leap in the fidelity of artificial intelligence (AI) art output has been made possible by advances in deep learning technologies, particularly natural language processing and adversarial generative networks.
Essentially, a user can enter a textual description and the algorithm automatically translates it into a coherent image.
MidJourney – or MJ as its avid users call it – is perhaps the most alluring technology for its pictorial production and poetic interactions. The charm starts from the first moment, with the “/imagine” command line prompt.
MidJourney founder David Holz said users find their text-to-image interactions to be a “deeply emotional experience” with the potential to be therapeutic. He said:
“A lot of great things are happening.”
MidJourney plays with genre and form, using existing principles that have long influenced media arts practice, such as non-linearity, repetition and remix, to mine the archive.
Holz suggested that the purpose of the algorithm is “to increase our imagination”.
My first requests for images were whimsical requests, nocturnal flights of fancy, gentle attempts to launch into the virtual world of spirits.
It turns out that my soulful prompts were surprisingly well suited to the algorithm’s default aesthetic.
The magic also hides in the algorithm. Ilya Sutskever, co-founder and chief scientist of OpenAI, describes the process as “transcendent beauty as a service”.
Artist and theorist Lev Manovich has poetically described his interactions with MidJourney as akin to working with a “memory machine”.
The recognition that it is a service but also a metaphysical experience is a new way of thinking about automation tools.
The technical process can be an imprecise science in which slippages and overlaps are inevitable. As Manovich acknowledges, MidJourney remixes:
something from real history and popular stereotypes – real acquaintances and fantasies. But don’t blame him, because we do the exact same thing, all the time.
The MidJourney Bot is hosted on the social platform Discord, creating a heady cascade of generative screen works.
It is an inherently communal experience. Image Stream also functions as a shared authoring site. If another user’s composition catches your eye, you can co-opt their prompt — or the image itself — and refine it to suit your own aesthetic preferences.
This collaborative remix is what makes the MidJourney Discord channel as much a social experiment as it is a scientific one.
My research into the darkening aesthetics of digital media means that I am somewhat predisposed to spotting dystopian visions. The MidJourney Discord channel is certainly an enticing rabbit hole for voyeurs of destruction.
Scary future cyborgs and post-nuclear wastelands seem to be strictly for the AI prompt engineer. I regularly see prompts citing the retro-futuristic nightmares of artists such as HR Giger and Zdzisław Beksiński and the cinematic tendencies of David Lynch and Andrei Tarkovsky.
As Bowie sang on the cyber-noir album Outside, itself a chronicle of art world depravity: “There’s no hell, like an old hell.”
Users are also finding ways to apply the technology in a moving image context. Notable efforts include a generative mode demo, morphing amoebas narrated by a synthetic David Attenboroughthe participatory story SALT_VERSE by Fabian Stelzer and the fascinating fractal film Monsters by Drew Medina.
The most significant assemblage I have encountered is SOLAR (Machine Drawn History of Humankind) by Gabriele Dente, along with a manifesto outlining the associated ethical and industrial implications of neural networks.
Digital tools have long been enablers of speed, dexterity and adaptability for designers and artists. Studio professionals in the MJ community are already finding efficiencies in their workflows.
A surprisingly beautiful example of the possibilities of design and conceptual ideation comes from architect and designer Cesare Battelli.
His “space-kangaroo” series is evocative of a way of thinking about conceptual design that combines aesthetics, functionality and fantasy.
Eryk Salvaggio described the technology of the DALL-E platform’s more photorealistic aspirations as “a kind of spiritual photography” evoking images filled with the ghosts and marks of past technologies: the fading image, the support in decomposition and the corrosive chemical. reaction.
This ability to piece together the past and embellish the result with techniques of capture and display and procedural degradation makes MidJourney particularly fertile ground for “authentic” gestures of the fabulous and the fake.
However, the contribution of this sudden rise in synthetic media to the glut of misinformation online is uncertain. How does the visual historical record accommodate its synthetic mirror?
We should also consider the evolutionary implications for language and computation. With the democratization of AI assistants, the field of human-computer interaction is rapidly evolving, and so are the inherent entanglements.
And so, tonight, as the city sleeps, I watch the stream and dream with the machine. I type in another text prompt and eagerly wait for my MidJourney Bot to conjure up its response. All the while, I wonder about the scope of the text in the code of the algorithm, and to what extent is it recoding me little by little?
Mitch Goodwin, Faculty of Arts, University of Melbourne
This article is republished from The Conversation under a Creative Commons license. Read the original article.