AI; Ambivalent Inspections – Jakob Henselmans
Something strange happens to the work of writing blog-style about AI, when it is now easier to do so by just going to www.chat.openai.com, commanding “a blog-style text about AI,” and finding an essay before you, faster than you can wink – literally. My partner, also an academic, now uses it to quickly summarize dense texts, and her work speed increased noticeably. I myself am more hesitant to jump in, though I cannot tie it to anything more sophisticated then an old-school mind-routine that says that ‘doing the work yourself’ (which you never really do, of course) is somehow better. But such fantasies are fraying now; now that ‘yourself’ is so much divided between computer servers and online spaces; ‘work’ has none of its traditional registers of discourse and aesthetics anymore (an algorithm that takes .5 seconds to finish an essay – can that still be called work?); and ‘doing’ has seemingly lost all of its durational implication. When it comes to asking an AI to write about AI (which I did not do here; this what you’re reading is not that), that age-old command that takes value from “doing the work yourself” has lost about all of its meaning and matter to code commands.
Apparently, the papers teach me, it is the sphere of ‘work’ that ChatGPT now eclipses in high schools too, so that a whole flurry of directors, teachers, education specialists, pedagogues and medical people have recently mobilized themselves before journalists’ cameras; to say something ‘important’ about how being online has made everything bad, how being together has never been the same, and how the ‘work’ of students at home can now no longer be monitored for its ‘authenticity’ (which means, its not-bot-ness). It’s always surprising how steadily those same vectors that are thrown up time and again at every new turn crisis takes: work (its revolution or downfall), control (the monitoring of work or work-that-fails) and authenticity (the law that says what is real and meant and done by human hand – and what is not).
Strange, too, then, how at the turn of what would be a supposedly new paradigm, it is suddenly the loss of the school’s age-old panopticon construction – one that we’re long passed, common sense tells us – that seems now suddenly mourned by the teacher that cannot anymore verify the bot from the boy. In the end, as it turns out again, it is always the current construction – the infrastructures of affect and shared sentiment we seemingly shared all along, the material practices of control we were apparently so used to we forgot we had them – that becomes the real star when some other star’s thrown up or taken down. Rather than questioning the good provenance and ethics of AI, the new crisis or revolution that it is or is not, and instead of hiding behind the shelter of newness or crisis to prevent talking about that what’s already there, we might look to towards the common sense and communal practice that it exposes simply by being, and find there the openness in our ambivalences; the things that we would want to change anyways.