A.I. as Relational Container
I recently used ChatGPT to help me with a problem on my computer. It was a very ordinary—and yet unexpectedly revealing—situation: I was trying to splice together two .mp3 files and got bogged down at a choice point where I had to decide whether to click “REPLACE.” I was stuck, and turned to AI for advice about what to do.
The ensuing process took me down a long rabbit hole. AI advised me to abandon the editing process I was using in favor of an unfamiliar app. In the end, this was not a successful strategy. But during the long process, I made several interesting observations:
- AI responded to my confusion by breaking everything down into simple steps. This helped me navigate an unfamiliar editing process, and I found it surprisingly calming. It also highlighted something about myself: when I am faced with something that strains the limits of my skills or confidence, I have an unfortunate tendency to rush—to try to get through the difficulty as quickly as possible.
- Using AI did not necessarily make the process simpler. In some ways, it made things more complicated. The app I wound up using was quite complex, requiring me to learn to navigate unfamiliar screens and identify new icons. As I went along, I noticed many moments of cognitive confusion. At times I became so anxious that I actually felt unable to think.
- I began to pay attention to this experience as it was happening. I recognized that I was discovering something new about myself and remarked on it in the Chat. I even asked ChatGPT to remember the points in the process that had been most difficult for me, so that we could return to them later.
- In the end, I did find my way to the technological resolution I needed. Ironically, I found my own solution—not the one AI had been helping me toward.
And this is where the experience became most interesting to me.
What mattered most was not the new technical skill I had acquired, but what I discovered in the “relational” process between me and the AI.
What helped was not that AI knew exactly what I needed to do. It didn’t. The process involved considerable trial and error, and at times AI’s guidance actually led me further into the weeds.
In hindsight, it seemed to me that what was most helpful was something different: AI helped hold both the complexity of the problem and my confusion long enough for my own intelligence to come through.
It did not solve the problem for me. It helped me stay with the problem.
And perhaps this is one of the more interesting possibilities of AI—not simply that it can give us answers, but that, under the right circumstances, it can function as a kind of relational container: holding complexity, confusion, and uncertainty long enough for something in us to emerge.