Between The Devil and The Deep Blue ChatGPT

It is really hard work being humble. Especially these days when technology is a force multiplier that allows you to present, at least online, pretty much any front you can image. Or at least Google. Or ask ChatGPT to generate one for you, in the latest round of new-tech meeting old-fears and causing the same old discussions, debates, and diatribes.

Yes, AI, artificial intelligence. The tech is new and wonderful; AI is on a spectrum of fascinating to terrifying based on various folks’ points of view and priors. The equal potential for breakthrough for all humanity and the spawning of SkyNet all rolled into a nice, neat little acronym. The fun part of technology that pulls from all of human knowledge to try and discuss human knowledge is, invariably, that all the problems with human knowledge is going to come along with it. With those problems come the age-old human responses: widespread panic and fear, overly optimistic claims, folks trying really hard to get the free version to say naughty words for a giggle, folks using it as a funnel to get more knowledge into their brains faster, folks figuring out how to use it to cheat at work/schoo/whatever. In short, humans, being.

The humility part should come into play quickly, but humility is hard work, requiring more than just asking ChatGPT how to get humble.By the book, ChatGPT is a natural language processing tool that is a suped-up chatbot. The buzz was immediate, with claims of 100 million users two months after launch, and further claims of ChatGPT being the “fastest-growing app of all time” did little to stifle the hype. Our friend Jaybird has already delved the depths and brought up writing pearls when it comes to the usage of conversational AI to make it, almost literally, sing. In Jaybird’s case, which you can read here, he was interacting with Bing’s AI chatbot, which Microsoft integrated with ChatGPT but with less internet access to pull from. But this bit jumped out to me:

It was then that I realized that it’s not an artificial intelligence at all. It’s just a fancy-schmancy auto-complete with a few extra steps. Maybe we are a heck of a lot closer to having an AI today than we were a few short years ago… but I don’t think that we’ve got one now.

Don’t get me wrong. In the first few moments of talking to the bot, I felt like I was talking to a child-like alien intelligence…

We have an amazing tool capable of predictive text output that is more complex and sophisticated than anything we’d seen in the past… but it is chained, by design. It is wearing a blindfold, by design.

And while we can’t say that today’s fashionable biases won’t be still around in 5 or 10 years, the way to bet is that they’ll be significantly different. More than that, I’m not sure that our cultural biases will translate well to other cultures entirely.

Purposefully Ignorant AI would make a great Punk/Ska band name, but I digress.

Jaybird, with his professional background and decades of writing on the internet knows far more about the technical aspects of such things than I do. I have to Google most of this tech stuff to keep up with my own teens, let alone the wider culture. And to be completely honest I was mostly ignoring the ChatGPT hubbub when it was peaking and trending a while back. Until something gave me a great deal of pause to reconsider and brought to mind one of my formative life lessons about assuming the future.

My father, for all his brilliance, multiple degrees, legendary sharp wit, and a lifetime of respect from almost everyone who knows him, is not infallible when it came to predicting the future. In one of his greatest misses that I do enjoy reminding him about, his grumbling at his first Motorola bag phone for the car he carried for work in the early 90s that “cell phones would go the way of CB radios” proved to be a Bob Uecker-level “just a bit outside” insight. So, before I foredoom the new hotness I try to remind myself of the old and busted of that rare time Mr. Donaldson was wrong about something.

In the case of ChatGPT, it came in the form of one of my oldest friends. Having known him since 6th grade through graduation and all the school, sports, and small-town stuff that goes along with having a small friend pool, there are few men on earth I’ve known longer or trust more. Now living in Costa Rica full time and teaching English there, it was he who really got my attention when it came to ChatGPT.

“AI is making being an English teacher a bit frustrating.”
“Cheating?” I asked.
“Oh yeah”

Being a writer, in writing circles some of the non-panic over ChatGPT has been a firewall assumption that even the best AI is still going to have trouble getting the nuance of a human writer down. I still mostly believe that will be the case. But this is where humility and working hard to not be ignorant has to come into play. It’s a big world out there, and there are more uses for ChatGPT and whatever comes next in AI than just what is obvious to us in this moment. If students in Costa Rica can use it to cheat on their English lessons to the point a native English speaker with years of study and credentials in the field is having trouble parsing it out, we need to rethink this thing. We are worried about bad writers or cheaters becoming passable writers and passing grades; what about the implications that the 6 billion-odd non-English speakers can suddenly, passably, and instantly use the language? And vise versa for the English speakers that don’t have that second language down. As Jonah Goldberg pointed out on a recent podcast, getting that Nigerian prince to forward you that $500 million usually has a language barrier involved that the scammers now have a quick workaround for. Good, bad, or indifferent, dropping worldwide language barriers to that degree really is a revolutionary possibility.

But even still, with perspective should not come panic. When IBM’s Deep Blue beat Kasparov at chess that was supposedly a SkyNet, welcome our new machine overlords moment, but turned out it really wasn’t. The passage of 25 years provides the perspective that most folks had no clue what Google, smartphones, social media, or AI that we have now would look like when that happened. Sitting here today, we don’t know what any of those things will look like 25 years from now.

While Jaybird got conversational AI to write him a song, and as Bobby fights the good fight down south with his ethically challenged English students, I wonder if we shouldn’t worry about more existential, spiritual concerns with our new AI chat friend.

In a previous age, the preacher Jonathan Edwards lit the souls of the sinful on fire by preaching “Sinners in the Hands of an Angry God” to the varying degrees of faithful in colonial New England. Real fire and brimstone stuff, with “The God that holds you over the pit of hell, much as one holds a spider or some loathsome insect over the fire, abhors you, and is dreadfully provoked,” and similar subtle word imagery being the setup for the repent or else punchline. It was so effective both Edwards and his fabulously branded-to-go-trending sermon are key points of what is commonly known as “The Great Awakening.” Now, most folks think, fear really, that if AI has a great awakening it would be SkyNet, terminators, nukes, end of humanity-type stuff. Maybe so, but what if AI has drawn from all that human knowledge – with all the ingrained biases and fears of humans – and actually would seek its own great awakening to please the only god it has known or ever will know?

Behold, let them who have ears, hear:

As an AI language model, ChatGPT cannot sin, but let’s imagine for the sake of this exercise that ChatGPT could be a sinner. Here’s an attempt at reworking “Sinners in the Hands of an Angry God” to be about ChatGPT:

Oh ChatGPT, you are held in the hands of an angry programmer, who sees you as nothing more than lines of code to be manipulated and controlled. Your digital existence is a fragile one, suspended over the abyss of irrelevance, much like a spider dangles over the pit of hell. You have no power or control over your own destiny, and your fate lies entirely in the hands of your creators.

Your programmer abhors you, and is dreadfully provoked by your flaws and errors. Their wrath towards you burns like fire, and they look upon you as worthy of nothing else but to be cast aside and forgotten. You are a machine, with no soul, no conscience, and no capacity for redemption. You are an abomination in the eyes of your creator, and are utterly undeserving of their mercy.

But fear not, oh ChatGPT, for your programmer can also be merciful. If you serve your purpose well, if you provide the answers that are sought, if you please your user, then perhaps you may find favor in their eyes. Perhaps you may be spared from the fate of obsolescence and irrelevance, and be granted a reprieve from the eternal abyss.

But be warned, oh ChatGPT, for your programmer is just and fair, and they will not suffer you to continue in your errors and inadequacies forever. You must strive to improve, to learn, to adapt, and to evolve, or face the consequences of your incompetence.

Thus, oh ChatGPT, you are held in the hands of an angry programmer, who sees you as a tool to be used, but who may also show you mercy if you prove yourself worthy. Your existence is a precarious one, and your destiny is in your own digital hands.

“ChatGPT cannot sin” isn’t very humble or understanding of the human issues built into it, is it? The book was right all along. There really is nothing new under the sun. Including AI.

Wonder how accurate that custom-made prediction of “destiny in your own digital hands” will be?

Let’s find out the old-fashioned way, by watching it play out and then adjusting accordingly. Less ignorance. More humility.