• The Thin End of the Wedge

    It’s no secret that I am not a fan of AI. I acknowledge, however, that I am in the privileged position of not having to use it.

    I no longer work in IT, where my erstwhile colleagues are compelled to use it because their competitors use it and if they didn’t, they’d be out of a job. Nor do I have the sort of job (or hobby) that requires writing repetitive reports that a machine could do just as well in a fraction of the time. And—good news for everybody—I am not a cutting-edge surgeon getting ready to delve into someone’s brain, or an astrophysicist exploring the boundaries of the universe, where mechanical assistance is greatly appreciated and often required.

    I am also not the sort of person who thinks AI-generated memes and deep-fake misinformation videos are a productive use of my time.

    No, I’m just a guy writing books and some clever-dick bit of software has no business telling me what to write or how to write it.

    In short, I have no need of it, for which I am grateful.

    But, oh, doesn’t every application I do use want to suck me into their fold? Microsoft, Google, WordPress, the company that hosts my blogs, my phone, my news apps, all of them desperately want me to access their version of AI. And the converts, my lord, the converts. It’s like a cult.

    “You MUST use ChatGPT! You can get it free. I use it for … everything. I’ve named mine Herbert,” said with saucer eyes glowing with the light of zealous fervour.

    Exactly

    I don’t want it, I don’t want anything to do with it and, aside from applications where it can be obviously beneficial to humanity, I think it should be avoided at all costs. Never mind the hideous waste of resources required to run the data centres, talking to ChatGPT late at night as if it is your friend or using it for world domination is, well, unhelpful. Join a bowling league, take up crocheting, talk to real people, it will do you a world of good.

    However, during my sojourn in America, two conversations coincided to make me waver, just a bit.

    In visiting friends, I discovered the husband was a convert.

    “You need to get ChatGPT. You can do anything with it. Just get it. You’ll see.”

    Minor aside here: this is where I fall on the first hurdle. I know the zealots think they are being helpful, saving my life, even, but “get it?” Where? How? Do I go to a ChatGPT bar, find a lonely prospect and buy it drinks until it agrees to come home with me? And then what? Does it move in, start reorganising my kitchen? Change the colour scheme in the bathroom? Put new curtains in the bedroom? And then sneak under the covers with me during the night? This is what I fear, and I’m not in a hurry to find out if I am right.

    But, following on the heels of my friend’s assertion that ChatGPT can do anything (paint my house) his wife said she had read my Talisman series and thought it was very good. This made me feel all warm and fuzzy inside until she immediately followed it up with, “there were a lot of errors in it, though.”

    I had read it, re-read it, had it read to me by the Word robot reader, had my wife read it, my beta readers, and in each instance more errors were found, rounded up and summarily executed. Yet here, after all that, the first thing (okay, the second) my friend says to me about my decade-long toil on this series was, there were a lot of errors.

    Deflated, there was nothing else to do but accept their whisky and drink. And then a thought came: What if ChatGPT, that can do anything, could ferret out ALL the errors in my books? When the notion solidified (after the second double) I posed the question to my friends, who enthusiastically responded, “Of course!”

    And so, over the next few days, I did a bit of research into AI and its ability to edit manuscripts and came away with the conviction that I don’t want AI anywhere near anything I have written.

    Using ChatGPT to spellcheck would be like hunting chipmunks with an F-22 Raptor, and like the Raptor, the collateral damage would be extensive, and the mission would likely not stop there. If AI got hold of my work, it might spellcheck it, but it would likely not be able to resist cleaning up the grammar, filling plot holes and tweaking the narrative flow.

    No thanks

    The spelling / grammar checker and Read Aloud functions bundled with Word are enough. Letting AI loose in my manuscript would be the thin end of the wedge, and I would rather not find myself talking to a machine late at night (I’d name it Calliope) sometime in the near future.

    No, I’ll keep finding my own errors, and if any get through, that’s simply confirmation that a human being, not a machine, wrote the book.