N.O.R.A. (Neural Open-Ended Response Algorithm) is text generation engine trained on my own surveillance data. It responds to Hannah Arendt's idea of agency as a combination of work, labour and action, establishing algorithms as active users of the built environment and examining their ability to engage with it. Through conversations with N.O.R.A., I got to interact with my “other” self, which lives in server parks that are more similar to military bunkers than traditional storage facilities. Because of the ethical complexities of data collection, I have decided to use my own information to train it. This not only gave me an insight into the terrifying accuracy of both GPT-3 and online surveillance, but also showed me a different side of machine learning; teaching as a form of care. Through creating a digital reflection of myself I had to look at my own actions through the lens of otherness, and I learned to see myself in a new light through this detachment. My project also highlighted the processes through which this information can be requested. The convoluted steps and overcomplicated resulting files are by design meant to discourage users from accessing their own data. How can any surveilling body be held accountable when the very act of questioning it is discouraged? The pretence of transparency, the disclosure of the way algorithms work but only through overly technical jargon so it can not be understood is all part of our planetaryscale shift towards an omnipotent surveillance machine. N.O.R.A. brings this immaterial and untouchable “other” self into a material reality where it can be interacted with.

aronweber@icloud.com