Writers are worried:
These concerns stem from the changes enveloping our creative industries; the introduction of generative AI tools able to handle parts of the art, design, and writing processes; and the potential for a world of endless media in which the consumer doubles as the creator.
With these changes come heightened anxiety over what next, uncertainty about our collective future, and worry from the wordsmiths responsible for sculpting the information we all consume.
We can dismiss these concerns as unfounded fears worn by “newbie” writers or those with poor writing skills. We can brush them off with the reassurance that ChatGPT is more advantageous than detrimental. Or, we can ask: why? Why are writers worried about generative AI tools, especially ChatGPT?
So far, I’ve found writers’ concerns to fall under two brackets: job safety and relevance of the craft. I’ve read long threads on why writers shouldn’t panic but have found nothing that seeks to understand these anxieties. In this piece, I want to walk through these deeper fears before exploring the mantra, adapt or die. I mean, isn’t it through understanding that we offer a meaningful way forward?
The Great Fears
1. Job Safety
In Scratch: Writers, Money, and the Art of Making a Living, Cheryl Strayed tells Manjula Martin that “There’s no other job in the world where you get your master’s degree in that field and you’re like, well, I might make zero or I might make $5 million.”
Not all writers acquire a master’s degree before pursuing the craft, but Strayed’s words encapsulate the fact that writing is not the most lucrative profession. In their 2022 UK Authors’ Earnings Report, the Authors’ Licensing & Collecting Society (ALCS) revealed a sustained fall in professional writers’ income, showing that median earnings from writing alone are now £7,000, a 60% decrease over the past fifteen years. In 2017, the median pay for full-time and part-time writers was $20,300 and $6,080 in the U.S., reflecting a 42% drop since 2009, according to The Authors Guild 2018 Author Income Survey. These low financial figures are not exclusive to book authors. Journalists in the UK have been reported to earn less than £20,000 a year, compared to media professionals in other fields like public relations. And don’t get me started on the employers who present content writers with $10 for a 20,000-word piece.
Of course, some writers earn well, even better than professionals in certain high-profile sectors. But that depends on the industry and other factors: portfolio, location, status, access, pricing style, etc. A good direct-response copywriter will likely earn better than a poet. Where the poet mostly relies on book purchases and royalties from the little earnings of poetry presses, the copywriter makes her money by having a direct impact on a company’s profit margin and receiving royalties on their sales.
On a broad scale, dipping your fingers in ink does not guarantee a high earning potential and writers often have to supplement their earnings with other sources of income. So, writers have valid reasons to worry about their jobs, especially when those jobs are threatened by generative AI tools able to write and illustrate a children’s book (however flawed) in under three days. If you have an industry where tools can provide you with the final product at 90% of the quantity for 10% of the costs, most professionals within that industry are bound to have it rough.
2. Relevance of the craft
I started writing professionally in 2020, not too long ago, but long enough to try different forms of writing and learn (from arguing with empty Google documents and working with good editors) that writing is a craft; a fusion of artistic expression and scientific design. While not all writers need worry about job safety, I think every writer should have some concern about the relevance of the craft.
More than ever, we’re in a world where the end product of writing (AKA content) is more valued than the craft itself. Where the art of words, the process of ordering and stringing ideas through writing, seems less important than the final work; and the interior benefits of writing are shelved beneath its metric-driven outcome. Most people don’t understand the work that goes into creating written material. Writing is often broadly boxed as typing, a simple activity anyone can perform. There’s some truth to that, of course. Writing is perhaps the most universal medium of communication after speech. Almost anyone can compose a text message, create social media captions, and enter prompts in a search engine. It’s all writing, a lot of which technology has simplified. So, it’s unsurprising to see the craft reduced to a single portion of the process.
But when we define writing as typing and place greater importance on the end product, we minimize the robustness of the process and silence the other benefits of writing that we can’t instantly touch or measure. Writing is thinking, and as Nathan Baschez observed in Divinations, it is also searching and testing and engineering and leading, all of which are advantageous to the self.
Also, Writing has an enduring significance. A written work — not the click-baited momentary kind — remains relevant long after its initial publication. But we don’t often see this because we tend to zoom more into its immediate benefits (the reaction elicited, leads generated, followers received), not its long-term impact (thinking clearly, building loyalty, advancing ideas); on its one-time effect, not the broader impact (its social, political, cultural, existential, spiritual significance).
ChatGPT and other generative AI tools diminish the value of the writing process and amplify our cultural emphasis on the end product. As language models grow in size, each new tool would likely improve in its ability to communicate ideas, recall information, and produce personalized content at scale. So, it does make sense for writers to be concerned about the relevance of their craft, troubled by the reaction of society to the next new way of championing content as king.
Regardless, it’s clear that the genie is out of the bottle and there’s no putting it back. Which brings us to ask: what’s to become of our creative work and how can we use ChatGPT (and other generative AI tools) to our advantage as writers? I’ll focus on the latter question in the rest of this article and attempt the former in a different piece exploring generative AI tools and creative work.
Forward Ever, Backward First
“It is almost banal to say so, yet it needs to be stressed continually: all is creation, all is change, all is flux, all is metamorphosis.” Henry Miller wrote in Sunday after the war.
History has shown us that life is restless, ever changing. This fluidity permeates every branch of our world, touching our personal lives, the natural environment, and even the position of companies within industries. We’re now witnessing a shift we don’t fully understand (yet) and it’s touching how we operate, the way our systems work.
But we also find in history that society eventually adapts to these changes. We recalibrate our paths as we enter uncertain eras. How do we do this? Maybe because we’ve learned that it’s either that or death, adapt or suffer. But I like to think it’s also because somewhere in our DNA are qualities that tools cannot replace. Qualities like empathy born from an understanding of this shaky human experience.
A good way to adapt to change is to, as Bruce Lee notably advised, “Be like water making its way through cracks. If nothing within you stays rigid, outward things will disclose themselves.” To embody this advice is to seek understanding. To seek understanding is to explore and question. That’s how we capture opportunities and make our way forward.
Facing the Limitations
Generative AI tools are capable of performing tasks way beyond anything we’ve known. But they do have some limitations that offer us opportunities.
Here are three major limitations worth noting:
1. Lack of empathy
Good writing is about knowing the function of your piece and the audience you’re writing to. It’s about asking: what do I want my words to do? Who will read the words I write? It’s about making good choices with words because you know that whatever is penned must be gotten right with language, and approached with care and purpose. The words must be ordered in a way that delivers the intended message. None of this is possible without empathy, the quality that helps us understand who/what we’re talking about and how to frame our words to feed the needs and emotions of an audience.
AI writing tools might pump out good sentences, but they have no emotional understanding of the human experience. How then can they convey empathy the way writers would? Empathy is a creative device for communicating data and information gathered through research. Without it, we can expect words written with little thought, sentences rich with claims but lacking in context, and articles that say something but offer nothing worthwhile.
2. Limited understanding of context
Context is vital in writing. It gives meaning and clarity to the intended message, prevents miscommunication, and cements a reader’s understanding of the text. ChatGPT has difficulty understanding and producing contextual results. It struggles with generating text relevant to a specific experience and appropriate for a particular culture. The results? Content that is confusing, offensive, or riddled with unnecessary information.
“You definitely need to sense check the results,” explains Content Strategist Kent J McDonald. “That’s why AI is not going to replace people entirely. For the subject matter I write about, you need someone with experience to know what is useful information relevant to the context.”
3. Falsity and lack of originality
OpenAI hasn’t quite detailed how they trained ChatGPT. But AI writing tools are trained on a dataset of text from existing materials (through the internet in ChatGPT’s case), and their performance is limited to the patterns and correlations present in that dataset. This feature results in inaccurate responses.
“My major limitation [with ChatGPT] is that the majority of what I got back is wrong. Not just out of date, made up. E.g. Slack’s number of active users – totally made up, not referenced. E.g. Cisco bought company ABC. No, that was Microsoft. Also, no source,” says Dominic Kent, Lead Marketer at UC Marketing, a limitation also observed by Freelance Writer Shreelekha Singh, who noted that “ChatGPT produces several false, made-up responses. For example, I asked for case studies for a piece on SaaS conversion rates and mentioned that each case study should have an original source. The tool came up with totally fictitious responses and broken links. I was shocked. Every case study had the same set of words with some different numbers.”
Additionally, ChatGPT’s dependence on a dataset of text hinders its ability to come up with new ideas or perspectives, and therefore produce original content in the same way that human creators can.
“I’ve used ChatGPT, free AI writers, and AI writing tools that cost $30K a year, they all pump out the same crap. I’ve even had one “trained” off my past writing, still wasn’t good. In my opinion, it has very limited use cases if you want good content. It can pump out generic stuff, which some companies may want. But overall, I haven’t found them to actually be good at writing,” noted Michael Keenan, Co-founder of Peak Freelance.
ChatGPT’s limitations offer us a chance to lean into our strengths as writers. We have an opportunity to focus on skills like editing and qualities like empathy, while (perhaps) assigning more technical tasks to tools like ChatGPT.
Facing the Future
“You write in order to change the world,” James Baldwin notably expressed, “knowing perfectly well that you probably can’t, but also knowing that literature is indispensable to the world. … The world changes according to the way people see it, and if you alter, even by a millimetre, the way a person looks or people look at reality, then you can change it.”
The works we read adjust our perspectives, shape our ideologies, and inform our choices. Despite the commodification of content and the low value placed on the craft, our society needs artists that can weave words in ways that help us reposition our lenses and move forward as a species.
It’s reasonable to worry about our craft and livelihoods as writers. I’d personally think it crazy if writers didn’t show a sliver of concern about generative AI tools. But the biggest consolation — which isn’t merely a consolation but a truth — is knowing that tools cannot replace the qualities that truly define us as writers. The writing process involves a lot more than typing. It’s an exercise of identifying (threads, needs, ideas), thinking, exploring, understanding, connecting with other humans, engineering sentences by ordering words with empathy, and presenting and evaluating the impact of the published material. You’re telling me a tool can do all of that? I don’t think so.
Another consolation is knowing that we can draw a line on our work and call attention to those who value our craft. As content marketing consultant Noelina Rissman put it, “On a general scale, this belief that anyone can write affects writers’ earning potential because (I perceive) the highest need of writers is from those with the mentality of “if I had time, I’d do it myself.” They don’t view writing as the intricate craft it is and therefore want to pay pennies. But those who want to move beyond servicing these sorts of prospects should connect with other writers in their industry. My experience has taught me that networking provides insights on pricing, upskilling, running a lucrative freelance business, etc. Because, at the end of the day, knowledge is power. And whether we like to admit it or not, a lot of our success comes from being influenced by those around us.”
Good writers aren’t going anywhere, but something will change about the writing profession. Perhaps we’re entering a future in which human-written pieces would hold more value, seeing as tools can generate content with speed. Maybe a booming industry of prompt writers will emerge as we shift into this future. We might even get personalized tools that churn out content using data from our individual speaking and writing styles. Regardless, one thing is clear: writers will use these tools in some way, to aid parts of the writing process. The question is, how?
I’ll close by exploring that question through the lens of other writers, showing how they’re employing ChatGPT in their writing process.
How Writers are Using ChatGPT
- Andrew Mayne on using ChatGPT for creative writing collaboration.
- Jennifer Lepp shares her experience with generative AI tools and using ChatGPT for titles and plots.
“I occasionally use ChatGPT to generate an outline or as a draft starter. It’s good for getting over the blank page syndrome, but I wouldn’t use it for my final draft. The magic is, after all, in the editing.”
—Content Strategist Kent J McDonald
- Alex Velinov shares a long list of use cases for Chat GPT in marketing.
- In his 3-2-1 Fridays newsletter, Dozie Anyaegbunam showcases ChatGPT’s use for creating analogies.
- As a freelance writer, Elna Cain uses ChatGPT for a list of tasks like generating blog outlines, meta descriptions, summaries, and more.
“I use ChatGPT for blog outlines on topics I’m assigning out to freelancers and for webinar titles. I’ve also used it to jumpstart my creative process when I’m stuck. It has always required a re-write, but at least it gets me typing and moving in the right direction. For anything else, I’ve felt it falls short. The copy is vague, the syntax lacks diversity, and it doesn’t connect ideas thoughtfully.”
—Kelly O’Halloran, Senior Content Marketing Manager at QuotaPath.
- Schlomo Genchin presents a detailed illustration of using ChatGPT to flesh out pain points, product benefits, and copy that creates an emotional appeal for a product.
- Filipa Canelas shares an amazing guide for writing LinkedIn hooks with ChatGPT.
“I use it when I have writer’s block. If a paragraph I’ve written just doesn’t look right and I can’t put my finger on it, I enter it into ChatGPT and ask it to rephrase. 10/10 I don’t use what it produces verbatim, but it usually dislodges the block by giving me a new workaround.”
—Grace Cartwright, Senior Content Specialist at Klaus.