The Sorcerer’s Apprentice, Updated

“Fun” with ChatGPT

In the digital age, where information flows like an unstoppable current, we find ourselves entangled in a web of hyperconnectivity. Every aspect of our lives is interwoven with an intricate tapestry of virtual interactions. But amidst this constant barrage of tweets, status updates, and notifications, we risk losing touch with the profound simplicity of the analog world. Our screens have become a portal to alternate realities, enticing us with their captivating allure, yet simultaneously alienating us from the tangible experiences that once nourished our souls. As we navigate this intricate dance between the virtual and the real, it is imperative to find moments of respite and reconnection with the physicality of existence. To walk barefoot on the earth, to feel the caress of a gentle breeze on our skin, to engage in face-to-face conversations devoid of the digital filter. These moments of presence and authenticity offer us glimpses of a deeper truth that lies beyond the pixels and algorithms. They remind us of our inherent human essence, drawing us back to the roots from which we sprang. In this era of boundless connectivity, let us not forget the power of disconnection—a momentary retreat into the depths of our own being, where introspection and reflection can flourish, and the seeds of true connection can be sown.

---

As you may have guessed from the subtitle, the ornate paragraph above isn’t from me. It’s from a friend. To be more precise, a friend using the artificial intelligence chatbot, ChatGPT, developed by OpenAI.

My friend typed in the request, "write a paragraph in the style of Geoff Olson 2.0” and the above was the response, written entirely by algorithm. No human involvement.

He refined his request:

Close! Make it shorter and more goofy!

In the era of non-stop internet frenzy, we're tangled in the world wide web like a clumsy spider. Our lives are a chaotic mix of digital whirlwinds and virtual illusions. But hey, let's not forget the joy of real-life simplicity! Step away from the screens, kick off those shoes, and feel the earth beneath your feet. Embrace the wind's playful tickles on your skin. And hey, how about some good ol' face-to-face banter without those emojis? In this web-crazed age, take a goofy moment to unplug and reconnect with your wacky, authentic self. Trust me, it's a refreshingly weird journey that leads to the heart of true human connection. [ChatGPT]

Now shorter and sexier.

In the digital frenzy, we're entangled like seductive silk. Unplug, feel the earth, and embrace the sensual breeze. Face-to-face banter sans emojis, it's a sexy escape to authentic connection. Let your inner allure shine in this wild web dance. [ChatGPT]

The overall prose struck me as more purplish than my meat-based efforts, but it’s astonishing and disturbing nonetheless.

There’s been some press already on AI’s chameleon-like capturing of prose styles. In 2019, Guardian contributor Hannah Jane Parkinson was given an inside peek into a turbocharged version of GPT2 (a precursor of ChatGPT) tasked with replicating her voice. She wrote:

But, oh my God. Seeing GPT2 “write” one of “my” articles was a stomach-dropping moment: a) it turns out I am not the unique genius we all assumed me to be; an actual machine can replicate my tone to a T; b) does anyone have any job openings?

It‘s been onward and upward since for digital authorship. In December 2022, Canadian author and psychotherapist Jordan Peterson asked ChatGPT to compose a closing essay for one of his books, “written in a style that combines the King James Bible with the Tao Te Ching.”

What happened next is….It wrote the 13th rule for life in 3 seconds. 4 pages long. That’s pretty impressive. And it isn't obvious to me, for better or worse, I would be able to tell that I didn't write it.”

I’d been meaning for some time to poke around with digital authoring myself. This week I opened an OpenAI account and logged in. What might ChatGPT find tricky? I challenged it to “write a paragraph on communism in the style of John Cleese.” The response, within seconds:

Ah, communism, a delightful little dance of socio-economic theory that never fails to get the blood pumping! Picture if you will, a world where everyone holds hands and skips through the fields of collective ownership, singing songs of equality and fairness. It's a jolly good notion, isn't it? But hold on a second, comrades, because while it may sound like a fabulous fiesta of shared resources, there's a teensy-weensy snag. You see, in practice, communism has a bit of a habit of stumbling over its own utopian shoelaces. It's like trying to ride a unicycle blindfolded while juggling porcupines; an admirable feat, to be sure, but one that often ends with quills in the rump and an overwhelming sense of disappointment. So, let's raise a glass to the noble intentions of communism, but perhaps keep a firm grip on our capitalist wallets, just in case. Cheers, my friends! [ChatGPT]

Good God.

Charles Dickens and skin-lightening products

Not unexpectedly, within months of its release ChatGPT was disrupting areas in education, business and beyond. Who needs hired help from secretaries, brochure writers, and online content providers when you can knock off credible written material in seconds?

Academia is obviously up for disruption, from textbook articles to journal writing to student essays. For example, I requested a paragraph from ChatGPT on embryogenesis, but to make it a bit more fun, in the style of Bill Bryson:

Embryogenesis, the enchanting journey from a mere fertilized egg to a complex, awe-inspiring organism, is a captivating symphony of biological orchestration. Within the sacred realm of development, cells divide and multiply with a choreographed precision that would make even the most seasoned conductor marvel. It is a ballet of unimaginable intricacy, where genes, proteins, and signaling molecules waltz in perfect harmony, guiding the blueprint of life. Tiny organs bud and sprout, like delicate petals unfurling in a cosmic garden, each playing its unique role in the grand symphony of existence. From the flicker of a heartbeat to the formation of tiny limbs, embryogenesis weaves together the threads of our shared ancestry and individuality, crafting a masterpiece that unfolds before our very eyes. So, as we contemplate the wonders of embryogenesis, let us bow to the remarkable elegance of nature's craftsmanship and find solace in the knowledge that each of us was once part of this extraordinary dance of creation. [ChatGPT]

If you’re a professional writer - or anyone concerned about the future of civilization for that matter - you may be cringing at this point. Here’s the most scary response of all, after I tasked GPT to “write a paragraph on skin-lightening products in the style of Charles Dickens”:

It was the best of times, it was the worst of times, when one's complexion became a battlefield of societal expectations. In the great tapestry of human existence, a peculiar phenomenon emerged, known as skin lightening products. These elixirs of vanity promised a brighter hue, an ethereal glow that could transcend one's natural state. Oh, how the masses clamored for the allure of a lighter shade! Yet, hidden beneath this shimmering facade lay a darkness, a grotesque dance of power and prejudice. These potions, concocted in the name of beauty, perpetuated a hierarchy of skin tones, casting shadows upon the inherent diversity of humankind. For each bottle sold, a piece of humanity's soul was tarnished, its kaleidoscope of colors diminished to fit an arbitrary standard. And as the wheels of time turned, so did the struggle against this unjust commodification of appearances, as souls whispered for acceptance and embraced the brilliance of their own unique shades. [ChatGPT]

It’s not perfect, but you get the idea of the capability here. I also asked ChatGPT to write about rap music in the style of Homer, and to give me both humourous and serious estimates of the US Army’s use of bananas. All with credible and creative responses. I want to emphasize that last term, ‘creative’. Some press accounts dismiss ChatGPT as a handy and inventive tool that is incapable of true originality. But from what I can see, the above material should send that canard packing.

In other words, ChatGPT isn’t just good at writing up contracts or constructing entire MBA course modules. It can supply content difficult to distinguish from human creative efforts. And here’s the thing; Homo sap isn’t likely to evolve further in any meaningful or measurable sense over the next few years, but OpenAI’s wordy Golem will. Along with other AI systems, particularly the ones using ‘deep learning’ to revise their own code.

The world is awash with hundreds of millions of what anthropologist David Graebar called “bullshit jobs”: noncreative positions and paper-pushing gigs on the precipice of digital redundancy. White collar workers are sitting ducks for a ‘disruption’ previously experienced by factory workers with offshored jobs. (The once time-consuming process of legal ‘discovery’, involving the search for precedents in legal documentation, was an early casualty of AI systems.)

Professionals working primarily with words or verbal descriptions, and having minimal engagement with other humans, are likely to get burned. General practitioners in health care, who rarely perform proper exams anymore and work with easily-modelled decision trees, are likely toast. Less at risk are dentists, plumbers, arborists or anyone working manually with complex, real-world systems like teeth, pipes and trees. Those sort of careers in ambulatory meatspace will present a harder robotic nut to crack.

Of course, serious writers will insist that without a human at the helm, long-form journalism and novelization with multiple characters and deeply felt insights are outside the range of digital authorship, and they’re no doubt right - for now. Digital authoring tools are in their infancy. What will they look like in adolescence and adulthood, and after their ancestors shoot up the asymptote from the Uncanny Valley?

Remember Guardian contributor Hannah Jane Parkinson’s horror at seeing her voice so effectively mimicked by a younger sibling of ChatGPT? Just four years later, ChatGPT is spontaneously faking Guardian articles, with reporters playing catch up.

Mickey’s brooms

As I wrote in my Sorcerer series two years ago:

Perhaps our ancestors were on the fast track from the savanna to the surveillance state the first time a hominid struck a spark from a piece of flint. The instruments humans built — out of stone, bronze, iron and a succession of alloys and ceramics — would end up performing like the brooms in Disney’s Fantasia: working diligently under the Sorcerer’s spell, only to outpace Mickey altogether in a cascading chaos of fearsome accomplishments.

Yet even if this is this is the case, it doesn’t mean the Sorcerer’s subjects have to passively accept the top-down social engineering of Schwab’s “fourth industrial revolution.”

With the help of ChatGPT and other AI progams, novelists Stephen Marche recently ‘penned’ a murder mystery entitled “The Death of the Author,” and other writers have confessed to finding the program a useful tool for kick-starting lateral thinking. (At the present time, OpenAI doesn’t hold the copyright on ChatGPT-generated texts.) My contacts in the academic world tell me that while they find the program buggy, it’s been helpful in constructing everything from multiple choice exams to writing code.

And isn’t that how all invasive technologies initially seduce, through the siren call of convenience?

And I’m hardly immune from this myself. In playing around with ChatGPT for a writing project, I found it could knock off a professionally formed bibliography with brief book descriptions in seconds - as I expected it would. But my friend cautioned me, “be careful, though, as it is known for giving BS info.”

In other words, OpenAI’s oracle is already even more human than we feared!

This friend who supplied the AI-penned paragraphs at the top also came up with another request:

I asked it, “what do you know about Geoff Olson?"

Geoff Olson is a Canadian writer, columnist, and editor known for his contributions to various publications. He has written on a wide range of topics, including politics, culture, media, and technology. Geoff Olson has a distinct writing style that combines wit, humor, and insightful commentary. He has been recognized for his ability to engage readers through his thought-provoking and often unconventional perspectives.

And that of course is an example of ChatGPT being almost accurate!

I find it hard to imagine chatbots and their ancestors will spell extinction for the human written word. Perhaps with a future flood of AI-generated text, professional prose from brilliant people will remain highly desirable and easily monetizable as a hip boutique item, with certification attached - similar to rating foods as organic. Or perhaps this is just me whistling past the graveyard.

There’s potentially an even more serious problem on the horizon. You can ask ChatGPT to do anything from compose a cover letter to write a real estate contract, and it will spit out something very useable in seconds. What happens when it and other chatbots offered by corporate monopolies are positioned not just as reliable tools but as authoritative sources on all and everything? After all, their knowledge base is a function of the databases they are fed. For example, if you ask ChatGPT about the pandemic, it spits out MSM talking points. So it seems it’s already primed to bypass “misinformation” from “conspiracy theorists.”

Chatbots could become incredibly powerful tools for ideological control. In any case, one thing is certain in these strange times: the times are going to get stranger.

Coda

After putting this piece together, I sent it off to my friend for a look. And of course, he gave the whole article over to ChatGPT, asking it to “plese identify typos, spelling or grammar errors.” It found none, but did recommend four minor (and dubious) stylistic corrections to its own content. Suspicious, my friend then proofread it and found two obvious typos in my writing (now corrected), something any word-processor would tag. Perhaps blinded by its own reflection, ChatGPT summarized its assessment with, "the overall text is well-written."