Most things don’t just happen. AI is no exception. Divorce for example, isn’t something that just happens overnight, there are years of small things that align to eventually cause a couple to call it quits. Similarly, mastering a skill takes time. From rock climbing to writing, greatness doesn’t usually happen in the blink of an eye. Like mastering a skill, or getting a divorce, AI didn’t just happen. There have been years AND years of small things that have aligned to get us to the point we currently find ourselves. From the early Turing tests on computing machinery, to Siri on our phones and Alexa in our houses, we’ve been slow walking our way to the here and now for quite some time. “Here and Now” of course is just an intermediate step along the slow walk to some “There and Then” of the indeterminate future.
Make no mistake though, a lot of things are going to start changing really fast.
The third iteration of ChatGPT was released in November of 2022. Five days later it had 1,000,000 users. To put this into perspective, it took Netflix three years to hit a million, it took Facebook ten months. This says as much about the speed at which word travels today as it does about the draw of new technology, but this technology is changing everything. Is that hyperbole? I’m not sure, but I’m sure ChatGPT could tell me.
So, what is it? ChatGPT has been described by NYT Kevin Roose as “the best AI Chatbot ever released to the general public.” It won The Atlantic’s breakthrough of the year for 2022, in which The Alantic’s Derek Thompson describes ChatGPT as part of “the generative-AI eruption” that “may change our mind about how we work, how we think, and what human creativity really is”.
Essentially ChatGPT is a response machine. You ask a question, make a request, look for information, and it gives you a response. I asked it to describe GPT, and here is the response that I got.
Generative Pre-trained Transformer, or GPT, is a type of language model developed by OpenAI. It is a machine learning model that has been trained on a large dataset of text, and it can be used to generate coherent and coherently written text. The GPT model is pre-trained, which means that it has already been trained on a large dataset and can be fine-tuned for specific tasks, such as language translation or text summarization. The GPT model uses a transformer architecture, which is a type of neural network that is particularly well-suited for processing sequential data such as natural language text. GPT has been very successful and has achieved state-of-the-art results on a number of natural language processing tasks.
You can ask it how to start an LLC, or write a play, a movie script, a poem, a novel, it can compose a cover letter addressed to a specific college or job, it can help you learn a language, or write code. You can ask for a climbing training regimen, or the best spots to rock climb on your next road trip. You can ask it to write a newspaper article, or a dialogue in the tone of Plato. Perhaps you want an idea for a potato chip commercial in the style of Hunter S. Thompson. It can draw up terms of a land contract that favor the tenet or the landlord. It can write an apology note, it can write a love letter. It can write a blogpost; it can give you a new recipe based on your dietary goals and restrictions. It can give you information on anything you could imagine, and it can do it all in iambic pentameter if that’s what you’re into. It can give you guidance on how to build a stock portfolio, or which car to buy for a family of four that likes to hike and has three large dogs. It can create a five-minute stand-up comedy routine based solely on rock-climbing. It can do all of this in a matter of seconds.
There are some obvious positives involved here. I just found three great cars that fit my lifestyle, read a terrible joke that included the punchline “that is why I hate rock climbers” and got a recipe for roasted chickpeas in a matter of seconds. I asked for a dialogue about basketball between Emerson and Nietzsche, the result wasn’t terrible.
This program is in the beta stage, and I would guess within a year or two the ability to compose works such as dialogues written in the style of Plato and Bukowski will be almost indistinguishable from Plato and Bukowski’s actual style, and tone.
There are some obvious negatives as well. Effects on academia are impossible to predict, but I gave ChatGPT one of the writing assignments I normally give to my Intro to Philosophy students, and the result would have scored an “A”. Daniel Herman of The Atlantic deemed ChatGPT “the end of High School English”. At the very least new methods of identifying plagiarism will have to be considered, and new ways of evaluating student work will have to be introduced. The ability to compose malware has become democratized. The capability to discriminate between better and worse writers through evaluating cover letters and applications will be nearly impossible. The list goes on and on, and surely educators and recruiters will have to adapt should they have the opportunity.
The AI sphere isn’t limited to chatbots of course. There are AI generators that can produce a video “person” who can speak as naturally as you or I when given a script. I can’t help but to imagine how this will change the world of on-line education. Imagine that a generated speaker who appeals perfectly to human students is given a script generated from ChatGPT based on nearly any topic we could imagine. That generated person can then appeal to and teach the student audience in a fashion that in many cases could be better than an actual human instructor. Questions could be asked and answered through ChatGPT 5.0 (just a theoretical version I am using for the example) which operates on audio cues.
A quick ChatGPT cost analysis of brick-and-mortar institutions of higher education vs. AI education along with a declining approval of liberal arts, an emphasis on tighter budgets mixed with political greed and strong lobbyists suggests that winds of change are blowing harder than ever.
It doesn’t stop at learning institutions; you could imagine learning anything through these means. A training program for climbers, a cooking program for chefs, a roofing program for builders. All without a single human doing any of the teaching.
Doesn’t someone have to write the code? Isn’t that the human element?
The code is written. The code itself, the AI now has the ability to write code.
None of this even begins to delve into the overlap between entertainment and art, or how this will affect the ways in which art is understood, crafted, and consumed. It feels like we are stepping into A Brave New World, and it feels a little uncomfortable. At the same time, I immediately want to ask ChatGPT to write about how Aldous Huxley might re-revisit New London using AI?
I’m not sure how this will all turn out. I’m not against artificial intelligence, in fact I think all intelligence is artificial, but I do think checks and balances are in order and needed quickly.
I would absolutely approve of a law that requires any product of AI to be stamped accordingly. The future works of AI that include movies, books, poems, etc, are not something I could see myself intentionally spending time on. Effort matters to me, struggle matters to me, the difference between AI Plato and human Plato matters to me, even if in every quantifiable way the AI is better.
I cannot say with any confidence what is to become of higher education, but I hope that effort, struggle, and the imperfect conditions of the human element continue to matter to others as well.