The world has been amazed of late by the advances shown in human-taught artificial intelligence that resulted in ChatGPT. For those few of you still unaware, it is a seamlessly interactive virtual assistant capable of near-perfect responses. It can interact conversationally, using the same questions and responses we would see from a reasonably intelligent human; from such prompts, it is capable of followup, discovering some of its own errors, and even challenging incorrect input.
The advances are astounding, the potential even more so — not merely for this application, but for the sister app InstructGPT and the architecture behind it. And yet, there are those who would warn us rather of its dangers, and of a humanity that is not yet ready for the tools becoming available.
One example is the college essay. Consider: Earlier methods of measuring academic progress involved simple answers, but were often easily evaded, as students came up with crib sheets and other methods of cheating. Essays were invented as a way to force cogent expression on a topic, the which can only be achieved with both mastery and confidence.
Except now, an AI can write it for you, making the method useless.
Some would say that colleges ought to move past that sort of metric, but how else can they certify the competence of (for example) journalism students — poli sci — English lit? How long will it be before Harlequin has its books written by A.I.? –And will anyone notice? (OK; perhaps that’s a bad example.)
What about the evening news? Outlets are already eliminating physical reporters, often sending a single person where before competing wire services would send three at minimum. What happens when these are no longer backed by a redundant staff, each an educated professional capable of some level of oversight and error-checking, but instead by a single A.I. that cannot itself be practicably overseen due to the lack of human involvement in the decision-making process?
How long is it before society loses the ability to communicate effectively without technical help, and instead relies on A.I. for everything from TV scripts to their own suicide notes, which courts will have to rule on whether they are genuine or planted by a murderer? Perhaps we will have A.I. judges capable of telling the difference… except then, the only winner will be the person with the most advanced model. Somehow, given the realities of municipal, state, and federal funding, I don’t see that being the courtroom.
Certainly, while there is a legitimate danger that people will be unemployed by automation, surely that’s a different discussion. And yet, in fictional treatments of the far future of humanity, where machines have been used to eliminate drudge work, the privileged classes engage in art, scholarship, and literary endeavor. When a machine can write usable copy at three cents per page (as opposed to the professional rate of twelve cents per word), it seems unreasonable to expect mere humans to compete.
What is the value of human life when a computer can create better?
Admittedly, the arguments I’ve presented aren’t as persuasive as they could be. They’re evocative, perhaps, but hardly convincing. Perhaps I ought to run this through an A.I., and ask it to punch it up a bit.
You can send cash to PayPal in order to help support us, set up a subscription donation at Patreon, or buy us a coffee. Remember: Today’s dollar will only be worth 87 cents tomorrow, so donate now while it’s still worth something!
Besides, you know if you don’t donate to us, you’d just end up piddling the money away on rent and groceries.