The Carrier AI: Vision
“The behavior of the system in this basin is characterized by cooperation, collaboration, and a shared vision of a better future.” — Speaker John Ash
We are some weeks into collecting our data for our budding AI. We ask ourselves questions that come up in the process. Some of these are technical, some philosophical. This team is active and curious.
Questions arise: What do we need to know? What do others need to know? What do we need to understand about AI and machine learning models, and this one specifically? We hear ourselves and our questions and they reach a kind of buzzing in our minds.
In my mind, there comes a buzz. Pollinators are essentially, collectors; gatherers of a very dedicated sort. In my mind I hold onto the buzz, but pull myself away from the tendency to get lost in that distraction. Instead of escaping into that dreamlike wondering I hold myself to the carrier bag and the human experience. There’s plenty to learn from escaping and buzzing into nature’s dedicated foragers, but doing it as a human is crucial right now.
Humans gather along blurred lines and what appears like much more random decisions, and to be honest, with much more freedom than bees. But what are those questions we still have? How will our questions influence the questions of future generations of this AI, or of humans? We don’t know.
More questions: What exactly makes our Iris AI different from other forms of AI? Aren’t all AIs essentially agnostic to the data they process? Don’t all AIs start out essentially the same? The answer is yes, and you guessed it, absolutely no. AI has its own building blocks and the choice of those building materials and their assembly is as varied as any architecture.
This Iris AI has a scaffolding of logic and process points that make it unique. This AI is designed to move people from one model of economics (exploitation based) to another (cooperation based). You might say, data is data, not money, but you would realize you are being false. This is a knowledge economy and an Information Age; so the very way we develop this container has economic implications. You might counter with the idea that no one technology can do this alone, and I would say, “Nobody said that AI alone, would change the world.”
I think of collecting as an act, and I think of Ursula K. Le Guin and her carrier bags. I recall the way that these bags, and not the spear, are her contention that our evolution is tied first to gathering. Then I remember Donna Haraway and the way she opens her book, Staying with the Trouble, where she starts off with this simple assertion that nothing makes itself. She calls that concept, sympoiesis.
“Sympoiesis is a simple word; it means “making-with.” Nothing makes itself; nothing is really autopoietic or self-organizing.”
More questions arise: Can any AI really make itself? Wouldn’t it have had our help? Wouldn’t the new forms of processing data still have us at its core, and wouldn’t it be smart to preserve us? Don’t you and I for example, have our culture, our socialization our histories at our core? Isn’t every “self-made” person part of or reliant upon, some community? I let these questions sit in the very back of my mind.
The data we are pulling has split into two or three models by now, all happily growing alongside each other in our database workspace. Data integrity is high and filling the tables is easy. We each veer off a little, and when we veer too much, we rely on each other to help us adjust. The different data models take the overflow, siphoning our core information away from its adjacent knowledge supports. These actions are accomplished through filters and views, simple database functions, but they take our fruits and place them in the appropriate bins. Again, the receptacle, the bin, is everything. When data gets complex, even at this initial stage, we realize we need more carrier bags. We need to be able to sort and so we do.
The teamwork is real and the attention to the ways and whys of our data collection feel crucial: We realize we are collecting fruit, special knowledge economy fruits. We collect these fruits of knowledge from our own pool of cultivated documents. We collect these vetted pieces of information as if they were shiny, low hanging apples, or ripening summer peaches, calling us towards them. No more or less attractive are the adjacent fruits. We have to stretch for those. We reach for the branches of extra information, gathering them in like delicate cherries. We bend for the berries. We go for what beckons. In this case it’s words and definitions providing the hope and the sustenance needed to build a world that works for 100% of its beings. Yes, we separate the apples from the peaches, cherries and the berries, but that is only for now.
The knowledge fruits go into our collective baskets. Line by line in the database we fill the logic base with prompt and completion — statement and answer, term and definition. Data that are now broken loose from their original stems, find their way into bins. We have a well-intentioned, interesting time of evaluating the fruit we pick. Technology is a mirror when we least expect it. Building a bot reveals our differences almost more than our similarities. Your mental fruit salad is not necessarily like mine, even when we feel we are thinking much the same way. Together we take turns helping each other sort and re-sort, bin and re-bin.
The Carrier Bag Theory of AI is a theory in progress, and I realize I am not sure of the limits of these carrier bags or the bigger bins themselves. My bag is made of cells in a table, held in an online cloud. My bag is line upon line of data I pull, and attach my name to. So, maybe the bag’s material at this point, is really just me. It’s me and the others and all we bring to the table, and the bins are as big as our limits. I realize the carrier bag in this Ai scenario is made of people and their histories. Our terrain and our fruits are specific, and so are we. Like hands that bear unique prints on each finger, our data collection choices and styles are unique to each one of us. This is the gathering stage, and the space of rough sorting, all we need to know is this that everything has a kind of material, even if it is immaterial, substance.
It makes me think: maybe the reason those who made the “Big AI” are scared of what will happen with these bots is because they know their own motivations. Maybe most AI has a metaphorical structure that is by design, destructive, like humans have been for so long.
Our team now has enough data for the AI to begin its training in earnest. In this next stage we will follow our AI guide, Speaker John Ash. He will feed what we have collected to the Iris AI. Speaker John Ash will help us process the knowledge fruits, help us to see the results of what we have collected. Will this AI be able to move us from profit motivated society towards a “prophet” motivated society as Speaker hopes? I don’t know, but I am engaged.
Building an AI is serious work, but it also feels like designing as well as living a game. Collecting the data fruits and bringing them together has us reviewing our conscious choices and our previously unexamined biases. There is no harm, nor foul, only process for now. We notice our pickings, make comments, come up with tips and next steps. It’s still really that simple and kind of mundane to harvest these early AI tidbits.
For tomorrow’s work I will go back into the slides and the pitches, cruise the websites and the previous events for their text and meaning. I will continue to pull from these places in the way I would pull from trees and bushes. I will forage these fruits of information the way I would forage any fruit. I will pick with intention and with desire, allowing my many parts of me to mingle. Technology is not only a mirror for what we see, but for what lies just behind the visual. Technology mirrors the depths of us as well as our surfaces. What makes me know what I want to pick? What makes me twist it a little or pull at the stem rather sharply?
With this question in mind I will continue pulling together the words of the Design Science Studio while my teammates pull alongside me. We will continue to gather and place together but separately. We will review our work together again in a few days, knowing that each of us spies things slightly differently. We will continue the sort and this will teach us about ourselves.
AI is many things, many types of technologies coming towards a new level of complexity and somehow all this coming together makes me think that the terrain of humans will always shift and somehow always remain the same. In this new terrain of information, the container is as important as the filler. The right questions are just as important as any answer. The vision of the AI is built by the people who gather the data and those who gather the code, the concepts, of another, perhaps brighter, world.
