Inspired by Bruno Latour’s Where Are the Missing Masses? The Sociology of a Few Mundane Facts, the AItball aims to question when and why we have delegated our decision making to technology.
Deliverable: Physical prototype
Timeline: Spring Quarter 2018 (10 weeks)
We met 10 times over the course of the Spring Quarter 2018, during the first six weeks we were tasked with a weekly reading assignment on design philosophies, some focusing on critical design. Based on these readings, we created a low-fidelity prototype to bring to class every week. In the last four weeks of the quarter we were tasked with taking one of those prototypes as far as close as we could to a final product. With the help of my professor, I reconstructed a Magic 8-ball with an accelerometer-equipped Arduino and LED screen.
The Syco-Seer and Humble Beginnings
The Magic 8-ball was conceptualized by Albert C. Carter in 1944 as the “Syco-Seer” and through multiple iterations has become the cultural artifact we know and love (and purchase one million of per year, according to Mattel). Originally created as an homage to the inventor’s fortune-teller mother, I saw it as a meta-prophecy of delegating personal decisions to the powers of the unknown.
After learning of the mystic context of the Magic 8-Ball, I was inspired to find modern-day parallels to the mystical nature of fortune-telling in Carter’s time. Algorithms seemed to be an apt equivalent. Whether you’re scrolling your social media timeline or playing chess against Deep Blue, algorithms are a black box of decisions, inaccessible to the public and sometimes hardly known by their engineers. In some ways, these algorithms are making decisions that we can’t control, very similar to how an 8-Ball makes choices for you. And thus, the concept of algorithmic mysticism was born.
Much like the Magic 8-Ball, the function of the Magic AItball is to make decisions for the user, except instead of being a biased form of Donald Duck’s flipism, the AItball informs its answers by pulling user data from web history and recording the users. Basically, it’s an uncomfortably knowledgeable device that makes decisions for you.
The original 8-Ball is built from an oversized billiards 8-ball with an embedded cylinder filled with blue-dyed alcohol and an icosahedron (20-sided die) with 10 affirmative, 5 non-committal, and 5 negative answers. With careful use of a hand-saw and clamp, my professor and I deconstructed my Amazon purchase to extract the cylinder containing the decisive icosahedron.
Following the deconstruction of the original device, my professor and I set to work on creating the code that would run the AItball’s decision making. (Shhh — it’s not actually AI) The code would be uploaded to an ADAFruit Featherboard with a small LED screen soldered to the board. The following if/else statement is the cornerstone of the code used to determine how answers are randomized and displayed on the face of the AItball. Embarrassingly, the most time consuming part of creating the AItball’s software was finding out that there are two kinds of USB cable–one for data transfer and power, and one solely for power. The two of us spent a couple hours scratching our heads while unwittingly using the cable meant for power while attempting to upload our code to the Featherboard.
The LED displays a modified version of the original 8-Ball’s answers (i.e. “My sources say no” becoming “Data not in support”). When the flips the AItball downwards, the device clears the current message. When the device is again flipped upward, a new message is displayed at random from the array of pre-programmed answers, like a digital coin flip. In a nod to its heritage, the Magic AItball is biased towards affirmative decisions rather than non-committal or negative decisions. If this software was developed to the extent of the concept, the AItball would not just choose random answers, but rather make informed, contextual decisions based on the user’s personal browsing and behavioral data.
The AItball is based on a concept from Bruno Latour’s Where Are the Missing Masses? The Sociology of a Few Mundane Facts, wherein he explored the morality present in objects, amongst other things. The gradient shown below is a representation of the abstraction of morality based on the physical form of something responsible for an action or behavior.
This abstraction of morality is explored by the AItball. It asks the question: What changes when humanity is removed from complex decision-making? When the device is making a decision, how does this change the relationship between device and operator?
When a device is “smart” enough and its creator no longer understands why it takes certain actions, it puts into question if the device can have intention. I believe intention cannot be present without some sense of morality, so ultimately this device has morality. Without falling too deep down the rabbit hole of AI sentience, I want the AItball to make the user question if they’re comfortable with not knowing how a decision is made and consider what the implications of that are.
This device begs the question: why did you delegate your decision-making to technology you know nothing about?
Food for Thought
- If this device is using your data to inform its decisions, are you responsible for its consequences?
- What morals are built into such a technology, does it have your morals or has it created its own?
- Where have you relegated your decisions to an algorithm?
I hope this project initiated some discussions — if you have thoughts you’d like to share, want to meet up for coffee, or you’re the long-lost descendant of Albert C. Carter, send me an email! Thanks for reading.