Color Blocks Live!, 2024
The current premise of AI is to facilitate the human decision-making process. Whether this is done through text or image, there is an initial conceptual prompt generated by the user, followed by a response of amalgamated and formatted information from across the internet. This creates a shortcut for web research or storyboarding, allowing the user to feasibly compress time.
The beauty of the internet is both its expanse and ability to connect. We know more about other people on this earth than we have ever before; yet at the same time, through that understanding, we have to reconcile an acknowledgement of extreme lifestyle differences. Aside from the physical and environmental variations, we have to account for the array of language translation across space and time. This includes both foreign translations to native tongues, as well as the poetic interpretation of words and sentences from a single language.
In this small-scale project I started with a specific intent and a vague command: I wanted to generate new color swatches from a black page. As AI is a collaborative endeavor, I hoped for a random solid color swatch output (at a minimum), and a logical progression of color variety in conceptual congruence (ideally). I took screenshots, denoting the settings, wording, and adjustments that progress throughout the project. The actual result was completely unexpected, which though OK for an exploratory art project, ultimately leaves the question of the baseline understanding through which we are currently ‘training’ artificial intelligence.
As children, we grow up with a full range of senses: our communication is completed through nuance due to the in-person experience. Beyond vocal tone, we become attuned to the smallest gestures and changes in disposition based on proximity alone. There is the notion that we “feel” energy or vibes from other people; That tension is not only the flexing of a muscle, but the conveyance of an emotion or reaction.
The information we put into machines does not guarantee comprehensibility on the part of the viewer, nor the machine. Just as a profile doesn’t denote the same information as a portrait. Understanding multifaceted perspectives within a framework of continuity takes an awareness of situational distinction. That a conversation can diverge into multiple branches, but the entire narrative is connected and construed through a relationship of understanding between the participants. This is generally determined through a series of interpersonal checks and balances, where the coherence of the conversation may not be as explicit as the black and white record of text; But are instead dependent on references that, while inferred by participants, are not as straightforward to an audience as a whole. Or conceptually and colloquially called ‘twin language’.
When examining how we can put our collaborative processes with artificial intelligence to the most effective use, we need to also examine how we can mitigate mistranslations. Or, to say, if we are going to rely on artificial intelligence as a permanent measure of the past and solid foundation for the future, then we need to understand the predictability and reproducibility of our results.
