Technology: The future's ... not all write?


09 March 2023
Now ChatGPT has arrived, is AI going to put us all out of a job? Gary Dalkin looks at the way recent developments in artificial intelligence might impact on writers.

Artificial intelligence has been a familiar concept at least since the talking computer in the original Star Trek (1966), and the malfunctioning HAL-9000 in 2001: A Space Odyssey in 1968, a ‘character’ written with what now looks like considerable prescience by Sir Arthur C. Clarke. Douglas Adams humanised AI with a depressive robot, Marvin the Paranoid Android, and an elevator which sulked in the bottom of the lift shaft in The Hitchhiker’s Guide to the Galaxy (1978), and the term AI itself has been in mainstream use at least since Steven Spielberg’s 2001 film of the same name.

For over decade now we’ve grown used to AIs like Alexa, Google Assistant and Siri in our homes and on our phones. But then last autumn new apps like Stable Diffusion and Midjourney suddenly meant that AI-generated art was everywhere (including on self-published book covers and comicbooks), and shortly afterwards came ChatGPT (Generative Pretrained Transformer). ChatGPT is either a chatbot or a natural language processing tool (if you want to be posh about it) driven by OpenAI’s GPT-3 family of what are known as large language models. 

But that’s enough tec jargon for one article. ChatGPT has caused a stir simply because of how (relatively) good it is. It can’t write a novel yet, but it can produce perfectly good advertising copy, or write a simple short story, at least in broad outline. I asked it to write me a marketing style description of my hometown, Bournemouth, and it did a good, if bland job. The copy lacked specifics, so I asked it to have another try, and this time to include three of Bournemouth’s best attractions, and the result was much better. It was clear to me from this that the job of copywriter was about to be consigned to history, and indeed just before I started to write this piece the Guardian published an article by Henry Williams titled ‘I’m a copywriter. I’m pretty sure artificial intelligence is going to take my job’. So naturally I did what anyone would do in my position — I asked ChatGPT, ‘What do you think will be the main challenge writers will face as a result of apps like yourself in the coming years?’

ChatGPT: One main challenge that writers may face as a result of using language models like myself is the potential for over-reliance on the technology. If writers become too reliant on language models for tasks such as proofreading, they may not fully develop their own editing skills. Additionally, if writers rely too heavily on language models for idea generation, they may not fully develop their own creativity. It’s important for writers to use language models as a tool to assist them in their work, rather than a replacement for their own abilities.

Now contrary to the impression the reply above might give, AI systems have no real-world understanding of the words they produce. Their ‘intelligence’ is truly artificial, no more than the rapid output of the application of a complex set of rules to a particular prompt. So while AIs can generate coherent sentences, there is no guarantee that those sentences will be factually accurate, or that any arguments it puts forward will hold up logically or even make basic sense. This was demonstrated in January when the website Futurism reported how CNET had been publishing AI-written stories since November, and The Verge noted that CNET had been forced to issue ‘corrections on 41 of the 77 stories the outlet published that were written using an AI tool.’ At the time of writing CNET were running the following disclaimer on all their AI-written content: ‘We are currently reviewing this story for accuracy. If we find errors, we will update and issue corrections.’ Editor in chief Connie Guglielmo felt it necessary to respond to critics with a story published on 25 January titled, ‘CNET Is Testing an AI Engine. Here’s What We’ve Learned, Mistakes and All’.

Nevertheless, the day is rapidly approaching, perhaps within a year or two, when an AI is able to write something that reads just like a real novel. Yet my sense is that no matter how convincing such a book is on the surface, it will never be more than creatively empty because there will be no character and no artistic or creative point of view behind the words. No real intelligence. It will be literally meaningless, though I suspect many readers won’t care if they can consume an endless supply of generic AI-generated e-books at 99p a time – or considerably less, as it will be a race to the bottom.

I’m reminded of George Orwell’s Nineteen Eighty-Four, so often lauded for its political insight that its qualities as science fiction are generally overlooked. But consider this passage, about the story’s heroine, Julia, first published in 1948:

Content continues after advertisements

…she worked, as he had guessed, on the novel-writing machines in the Fiction Department. …She could describe the whole process of composing a novel, from the general directive issued by the Planning Committee down to the final touching-up by the Rewrite Squad. But she was not interested in the final product. She “didn’t much care for reading,” she said. Books were just a commodity that had to be produced, like jam or bootlaces.

It was time now to look on the bright side, so I asked ChatGPT: ‘What do you think will be the main advantage of apps like yourself for writers in the coming years?’

ChatGPT: One main advantage for writers in the coming years will be the ability to use language models like myself to assist with tasks such as proofreading, research, and idea generation. These tools can save writers time and improve the efficiency of the writing process, allowing them to focus on the creative aspects of their work. Additionally, language models can also help writers improve their grammar, spelling and punctuation.

So there you have it. AI writing is already being published and it is up to writers to make the adjustment, just as we have with every new technology from the printing press to the typewriter to the word processor to the internet. Fortunately for us AI can’t create anything genuinely new or original, and still needs humans to administer quality control. Consider that every story outline I asked it to produce was obviously derivative; I asked for plots for realistic science fiction stories set in 2033, and it responded with transparently obvious variations on The Matrix and Total Recall. It was drawing not even on the best of current SF literature, but on Hollywood ideas already decades out of date. Writers can’t ignore the coming AI revolution, but, at least for now, original storytelling remains firmly in the domain of the human imagination.

You can try ChatGPT at:


This article first appeared in the April 2023 issue of Writing Magazine. You can buy it here: