In sum, I don’t know how you’re using AI, or whether you can relate to this experience, but as for me, at least, I cannot shake the feeling that we are not only conjuring a superintelligence into being; we are also coming to share, ourselves, as individual human beings, in that same superintelligence.
A friend of ours forwarded a link to an article about a recent MIT study that used brain scans to show that heavy use of LLMs like ChatGPT “reduces memory, critical thinking, and brain activity.” I sighed and marveled at the fact that it’s only been fifteen years since midwits were enamored of Nicholas Carr explaining that Google searches and hyperlinks were actually making them dumber. Ironically, that may have been correct in a way he didn’t intend: no one learned anything from that hyperbole, so here we go again.
Smith, a cognitively-privileged philosophy professor (it’s okay, he’s one of the good ones), reports what others of us have also discovered: as with most technological tools, you get what you give:
The experience I am reporting, I’m aware, is by no means universal. AI is making some of us a lot smarter, as individuals. But there’s a paradox here: on the whole it is making society a lot dumber. How do we make sense of this?
The answer has at least something to do with age. Those of us who are old enough to have learned to do research prior to the rise of the online search typically bring to our exchanges with LLMs a mature ability to scrutinize their claims, and, when in doubt, to verify these claims independently. When the exchange is shaped by such scrutiny, it offers, as I think I’ve illustrated with the example of studying an “obscure” language, a more fruitful, intense, and rapid learning experience than has ever been had in the entire previous history of humanity.
…It seems, however, that if you had not already oriented yourself in the world as “aspirationally omniscient” prior to the arrival of our new information tools — if, that is, you have no preexisting personal project of encyclopedism to which to strap your new booster rockets, then AI does not so much supercharge your own effort, as simply take off without you. You have to want to absorb, to internalize, to make yours, all of the flow of information between you and your AI study partner if you want it to transform you in any significant way, rather than simply to do your work in your place.
I fear that a near-equivalent way of saying the same thing is that you have to want to become one with the machine, if you wish to claim your human share of our new transhuman superintelligence.
I’ve been using ChatGPT to help find more of the music and writing I like. It’s like a more-refined Google search. I’ve used it to research questions which would have previously admitted of no known interlocutor. I grew up in a time where the only thing to do with questions like that was ask them of a parent or teacher, or maybe consult a reference book or a librarian. Otherwise, you could shrug and consign the answer to the “Here there be dragons” section of your mental map. For people like me, the internet and chatbots have been a godsend, allowing me to read and learn things that would have remained forever unknown had I been born a few decades earlier. Nietzsche knows what I’m talking about:
“Oh, my greed! There is no selflessness in my soul but only an all-coveting self that would like to appropriate many individuals as so many additional pairs and eyes and hands – a self that would like to bring back the whole past, too, and that will not lose anything that it could possibly possess. Oh, my greed is a flame! Oh, that I might be reborn in a hundred beings!” – Whoever does not know this sigh from firsthand experience does not know the passion of the search for knowledge.
Of course, I’m sure that Oliver Bateman’s dystopian vision of the first human-AI marriage will become real sooner rather than later. And I’m sure that the web will continue to be awash in obvious AI-written articles passed off by humans too inept to make them sound unique. Technology didn’t create sad, broken humans, though, nor did it teach them how to cheat. Technology’s sin is that it confronts us with the results of our choices and reveals our priorities, and for a species as enmeshed in fantasy and self-deception as we are, that’s unforgivable.