Extended Intelligences II¶
I started the Master’s thinking AI mostly as a machine to get answers and with each course releated, i think more as a tool. Apart from learning about the technical side of how things really work, which is still important to learn to grasp the reality and the potential, we got an overview of what is possible and maybe what’s coming next. Taking the artifical intelligence from a passive element and to use it as an (active) agent that works with local data to give personal outputs was felt like a potential future usage. Eventhough Pietro helped us a great deal, i left the workshop(?) feeling capable; being able to make things that i don’t know how to do. This brings me my favorite part of learning; learning how to learn. For that, thank you Pietro & Chris!
The Fortune Teller¶
We started the ideation with thinking about how would it be to gossip with AI. In most of the courses about AI, we learn how it is just a translator of a human mindset, and that is not a black mysterious box. In our brief project, we wanted to bring more mystery into it. From our prompt to our main theme it was all about the mystery and oracles. In Turkish culture we have a ritual where we (most of us) do for fun after drinking coffee, that is fortune telling from your remains of the coffee in the cup. This probably goes way back, and a lot of cultures have similar fortune telling rituals; one with the hand lines, the other with moles in your body… In the porject we initally wanted to work with the face and use a biased algorithm to generate a fortune telling using that biased information. This we thought would be a nice showcase of how police and government are judging and labeling people. However, we understand that this would take more than 2 days soon (thanks to Chris) and as a proof of convept we used a more managable sensor; temperature and humidty. As a said before, i had no idea about how to create an AI agent but at the end we managed to create a AI agent using Open AI and Cursor to come up with code and print the fortune telling with a thermal printer! Working with Raspberry Pi made AI tangible; a machine with temperature, latency, and failure, rather than an abstract cloud intelligence.
What we created with the agent and RaspberryP Pico was no different than what a human fortune teller would do; create a story using something about that person with a mysterious delivery. AI doesn’t reveal truth; it reveals how easily we accept generated meaning when it’s framed as intelligence and felt personalized.
The part that stuck with me the most about the course is the discussion we had about using (hyper)local data. Hyper-local data is not valuable because it’s accurate, it’s valuable because it’s situated, embodied, and personal. So in the end we didn’t build a machine that predicts the future, but i believe we built something that reflects how futures are constructed. I stopped seeing AI as a neutral tool for problem-solving, not that it’s ever neutral, and started experiencing it as a system that actively constructs meaning, belief, and authority. This project made me question not what AI knows, but why we trust it.