I have a confession to make. It’s something that I as a writer feel morally conflicted about, a little bit shopsoiled even. Forgive me father for I have sinned, but please don’t shoot the messenger. My last two blogs were co-written with ChatGPT. There, I’ve said it.
Now, remember, that’s co-written meaning that I still directed, edited and included personal anecdotes and observations making them mine, but still leaving the machine to do some initial and remarkably fast heavier lifting. These were prompted from OpenAI’s oracular free version of its LLM (and we are not even talking its new cognitive version o1).
And think about it, we’re just one month short of the second year of its release to the public in November 2022, if you recall. How far it, and others of its ilk have come along in that time. Disruptive is just the beginning and the adoption curve compared to other communications miracles like radio, TV, WWW, digital and social media is three miles high and rising. The only thing standing in the way of bigger is the megawattage it takes to operate and evolve these AI lightbulbs that burn ever brighter.
After years of hype over tricky algorithms and chatbot personae masquerading as intelligence, we start to get something resembling an oracle, not a perfect one but one which draws on an expanding corpus of verbal, visual and videographic data out there to offer us coherent and complex responses– the answers we sought delivered by a Janus faced god sifting through all of our utterances to offer up a new mirror to ourselves.
We are its subject, and like anything that’s free and always will be, your data becomes your votive offering instead and feeds the hungry, highly observant beast. And remember, we know it’s dangerous to dally with the creation of an intelligence higher than our own but that’s the temptation of artifice. It’s attractive. But at the same time, it’s so enticing that it’s almost malign in its enticement.
Yuval Noah Harari prefers to characterize GenAI as Alien rather than Artificial Intelligence because despite its database being humanity, its instincts (if one can talk of such things about machines) are purely technological. It might be programmed to react with emotions but just like in Blade Runner, the memories that form these emotions have been implanted into the skin jobs — indistinguishable from humans, but manufactured, not born.
When I was writing this blog, I toyed with the idea of entitling it ‘incantation’ because there is something quasi-mystical about this new technology — something indistinguishable from magic. Whether it’s just language automation like a magician’s sleight of hand, or whether it really does what it says and is offering us true intelligence and insight, we are quite literally spelling spells and summoning up knowledge like a magician might summon a demon. It’s a new level of uncanny but it’s also a dangerous democratization of super powers.
When I came to think of it, from a narrow technological plane, LLMs have many similarities to the oldest extant work of literature developed by the Chinese four millennia ago, the I Ching. The Book of Change as it is also known, is something I have enjoyed working with for years. Whether you believe in divination or not, if you have ever consulted the book, you become aware of a higher intelligence guiding your thoughts and actions. Its answers are both time-bound and yet universal.
This then, the creation of oracles seems to be a natural human response to the unknown and seemingly unpredictable, to create systems of knowledge to better understand it. It’s the cornerstone of all learning, Of course, LLMs are mechanical brides divorced from nature, (upon which much of the I Ching’s incantatory beauty lies) but they are already proving themselves to highly adaptable, which as Darwin pointed out, one of the key success factors of any species.
Writers, and all other creatives were never meant to become sorcerers ‘apprentices, but in the first quarter of the 21st Century, here we are. And you know it’s only going to get weirder — like not when it upends whole industries that were originally ring fenced — stuff we take as part of the scenery of modern life, like film and publishing, but more concerningly, when it makes a whole white-collar class redundant (the same ones currently and diligently going to their jobs most days). That’s hugely impactful.
Sensing the writing on the wall, Sam Altman, OpenAI’s founder posited that governments should step in to pay everyone universal credit while we figure out what to do with the lately (or soon to be) disenfranchised. It’s an unfixed problem for the near future, and at the speed this technology is now advancing, it won’t be fixed in time either.
On a societal level, already polarizing between the richest and the poorest, the technology tilts the odds implacably in favour of a small technocratic ruling class and their sponsors (some good but mostly bad actors) while everyone else goes hang. I hope that all the talk of new disruptive technologies leading to new kinds of jobs will now be consigned to the ‘busted’ category in the same way that we know that trickledown economics, widely sold to the poor as their just rewards for serving their overlords has been proven to be just more shit rolling downhill (as usual).
One thing occurs to me though, and it goes back to my original confession at the start of this blog, do you reject this advance in communications as the basket case Luddites did in their day to smash automated weaving machines, or do you get with the program?
It’s probably too late for me to start a new profession as a prompt writer at 60. And honestly, who came up with that term anyway? It sounds suspiciously generic, and not a little dull and fails to capture the magic in the bottle that good writing can deliver. Prompting alone is not writing.
No, I wish to retain the same title of writer that I set out upon at the start of my career but with a caveat, that I will accept its massive collaborative and automating power. Writers are good at asking themselves the right questions and framing questions for others — what does it matter then if the recipient of the questions in this case is non-human? Let the audience decide for themselves if articles written in collaboration with intelligent machines are authentic and worthwhile, or just cunning simulations thereof. At their best, they will match us and possibly exceed us. Maybe.
Career paths and greater or lesser literature aside, I am more concerned about our ability to distinguish between artifice and the real thing and for our love of technology to run away to the extent where we lose our own humanity. There is a common thread out there that we are living in a simulation and deep fake perfectibility takes us one step closer to that. We are necessarily courting an unpredictable force.
In that regard, I am reminded of the highly prescient science fiction movie, Forbidden Planet in which a raging id monster is created when a scientist called Morbius interacts with an alien technology found on said forbidden planet. The scientist barely survives, but his intellectual capacity has doubled but so too has his fear of the monster he has created. Interestingly, powering this deadly monster stalking the cast of the film is an underground machine said to run with the energy of 9,200 thermonuclear reactors — modern artificial intelligence like alien intelligence is as we know, resource hungry.
When Morbius is told he must share this alien technology with Earth, he refuses, saying, “Humanity is not yet ready to receive such limitless power.”
For the tech community, now in charge of this new beast, it’s always good to remember the adage that with great power comes great responsibility. Something so profoundly transformative cannot go unregulated. On that note I will leave you with a TikTok link to a very interesting talk given by a former Google CEO on what may possibly come next — people get ready and remember to be careful what you ask your oracle for.
And one more thing before I leave you, this is a fully human communication — no LLMs went into its creation.
