I’ve been of the opinion lately that, if they weren’t born into privilege, every writer in the third world needs to quit their job. 

On my end, that entails giving up a 15-year career that’s given me everything I could’ve possibly hoped for: close friends, a remarkable portfolio, a diverse network of clients, and several lifetimes’ worth of memories, from staging sex comedies at the CCP to getting a grant for a childhood dream. At the same time, however, all that experience includes having been on several sinking ships. I feel like I’ve gotten fairly good at spotting one that’s about to go under.These days, that ship’s name is Writing Itself, and AI’s got an eye on the hull and a drill in hand.

How Technology’s Kept Writers’ Wages Down

What most people don’t realize about writing as a profession is that most of the work out there isn’t all that creative to begin with. The businesses that hire us live or die by their ability to get noticed, and the most effective way to reach consumers is to make sure your content shows up on Facebook feeds, on the first page of Google search results, on the endless scroll of TikTok’s For You page. And it’s algorithms—not readers—that determine which pieces of content get those coveted spots.

Since bots don’t have feelings, or brains to tickle, or souls to enrich, most freelance writers make a living settling for subpar work. Caring for your craft doesn’t mesh with the efficiency required for this sort of mechanical writing. Just write according to a predetermined metric, do it quickly, and get paid. It’s not ideal, but it works.

Advances in technology, however, often throw a wrench into that system. 

When article spinning hit the scene, for example, it created a ripple effect that kept writers’ wages down. The tech allowed our clients to generate multiple unique articles out of a single “seed article”. It automatically substituted phrases within the seed piece with similar ones, as long as you wrote in the proper syntax. You could, for instance, rewrite “I love my dog” like so:

I [love/adore/can’t live without]
my [dog/canine companion/four-legged friend].

Passing it through an article spinner allowed you to generate 9 unique sentences, each one using a different combination of the words in the brackets. And since much of your success with Google’s algorithm depended on having mass quantities of unique content, spinners saved businesses a ton of money.

Instead of hiring nine writers to produce nine different articles on the same topic, clients commissioned just one article and spun it. This drastically reduced the number of available jobs, and writers undercut each other by lowering their rates; since quality was hardly a concern, the cheapest writer got the job. Because of that, you can still find people in 2023 offering the same rates we saw on freelancer job sites in 2008: a paltry one peso per word. 

Today, you can come across videos teaching business owners to use free AI tools to automate the tasks they’d usually pay a team of freelancers for. “Here’s the secret to saving thousands of dollars a month,” they’ll say, referring you to apps that use AI to create art, videos, social posts, and articles. Instead of hiring one writer out of thousands that need work, the savvy entrepreneur just needs to pop a prompt into free AI writing software. If the pattern holds, grunt writers’ rates will continue to be unsustainable.

Given how quickly AI writing is developing, the smart thing to do would be to stop resisting the machine; to get out of its path and step onto one where I won’t get run over. It would be perfectly logical to just quit writing.

But for people like me, the act of creating is terribly addictive. There are countless ways to make much more money, but going too long into withdrawal from writing turns us into shapeless, desireless husks. 

So what recourse is left to us, in the face of what feels like an AI apocalypse?

How Can Writers Even Survive AI?

The obvious solution would be to focus on amplifying the sense of humanity in our work, to create things that AI simply can’t. But if the point of developing AI is to make it feel as human as possible, the definition of “humanity” as it pertains to writing gets pretty muddy.

Timothy James Dimacali, one of the Philippines’ foremost science communicators, recently spent a few weeks experimenting with ChatGPT, the AI writing tool du jour. I asked him what sort of impressions the experience left on him.  

“I don’t know if it’s enough to pass a Turing test, but it certainly feels very natural to me, to the point that you wouldn’t immediately think it wasn’t a person that wrote these things,” he shared in a mix of English and Tagalog. A Turing test measures a machine’s ability to produce behavior that cannot be distinguished from that of a human’s.

Despite acknowledging the progress that’s been made in making ChatGPT sound startlingly human, Dimacali believes that there will still be space for human writers in the future.

“My thoughts are based on the idea of writing, specifically–or any communication in general–as a discourse between an entity and another entity. And right now, one of the constants in this equation is people.”

“A lot of the fear that I think that’s being talked about these days is writers losing jobs, artists being passed over for AI, because why not?” he says. “It’s a whole argument for automation that’s been around since the time of the industrial revolution.”

Dimacali brings up how the invention of photography created a similar sense of panic among artists, and yet to this day, paintings sell for millions in a market that never seems to run short of buyers. What matters more than photorealism, he says, is the impact an image can have on a person. And the same is true for writing.

“Regardless of whether or not AI wrote the piece we’re discussing, people are thinking about how they’re affected by that particular piece of writing, which I think is the more important thing to take away from all of that,” he says.

“Even if TV scripts or whatever, or corporate writing–all of that sort–even if it was written by AI, it’s humans who are reading it.”

To this end, Dimacali sees opportunity in more reflective writing. Critique, for instance, is something he doesn’t see AI capable of producing, since it can only be written with a measure of introspection. His experiments with ChatGPT have also opened his eyes to another new industry altogether: what he calls “AI Whispering”.

“[AI] isn’t like a genie you can tell, ‘I wish for this,’ and suddenly you’ve got a complete piece,” he says. “There’s still some work involved. You still have to understand how to talk to the AI, how to prompt it in such a way that it gives you exactly what you want. And I think there’s something there.”

Even then, Dimacali struggles to qualify AI-generated output as the work of a writer. While the human crafts the prompts, it’s still the machine that creates the piece. 

AI researcher and poet Sasha Stiles sees things differently. Stiles uses an AI language model to write poems, which are arguably some of the purest expressions of the human experience. In order to do so, she’s painstakingly trained her AI to be more representative of her voice, essentially creating an extension of herself.

Because the AI takes in data and feedback from millions of people around the world, it becomes more of an amalgamated intelligence rather than an artificial one.

“I’m able to actually dip in and then hear very interesting and uncanny echoes of my own writing, of my own text, and of all the inputs that I’ve actually gone in and, in a very qualitative way, have trained this model on,” she shares. “It’s made it feel more personal and more human.”

Rather than a crutch, Stiles sees AI as a way to more fully explore one’s creativity. It isn’t too different, she points out, from Surrealists and Dadaists using their own aleatory writing techniques to temporarily detach themselves from the mechanical act of writing in order to discover new means of self-expression.

“It’s almost a way of countering some of the poetic pre-programming that I think I’ve brought to the table, and it’s enabled me to sort of feel more liberated with a lot of my writing,” she says.

“I think there’s lots of ways these tools are very efficiently automating some of the stuff that really is tedious about the kind of copywriting that there’s such a demand for. And in some ways, it really just enables you again to put more focus on concept and strategy, and thinking about an overarching vision for something, and being able to really put your emphasis as a creative person and a creative force.“

She posits that AI language models like ChatGPT, which rely on being trained by vast quantities of written examples rather than a single person, allow writers to tap into humanity’s collective consciousness. Because the AI takes in data and feedback from millions of people around the world, it becomes more of an amalgamated intelligence rather than an artificial one. And this, according to Stiles, could help grunt writers like myself develop more insightful voices, without losing the efficiency their clients require.

“Ultimately, it really is about synthesizing all this information from all over the place and helping us make sense of it,” she says.

What the Future Looks Like for Writers

It looks like the way forward, then, is to recognize that the more mechanical writing jobs, such as SEO writing, product descriptions, and captioning, will require writers to either learn how to collaborate with AI tools, or deal with the risks of a severely diminished client pool. There may still be niche demand for solely human-produced output of this sort, in the same way demand exists for artisanal goods and for paintings in a world with increasingly sophisticated cameras.

On the creative side, AI writing tools could actually help us be more in tune with the rest of humanity. Working with this “amalgamated intelligence” could, for example, create new opportunities in fields of insightful mass communication, such as advertising. We could conceivably discover emotional touchpoints we couldn’t have reached without AI assistance, leading to much more human messaging.

And for writers like me, for whom grunt writing serves to keep the lights on just enough for us to occasionally dip our toes into more creative pursuits, there still remains a lot of uncertainty. We don’t know how quickly AI writing will evolve in the coming months, or if we’ll be out of jobs sooner than we think. We aren’t at a comfortable enough income level to hit pause on work and really think about where we want to go. The stagnation of writers’ wages in the third world has robbed us of that agility.

The best we can do, really, is to try to step away from the “whatever works” mindset that grunt writing has ingrained in us and focus instead on developing our unique voices. If human-produced commercial writing will eventually evolve into an artisan’s market, then we need to put everything we have into developing our craft. Instead of writing for hits, we need to write things that hit different.

As an epilogue to this piece, I thought it would be cute to use ChatGPT for the first time in my life and ask it the questions that I’d hoped to have answered in the course of writing nearly 2000 words.

It took all of 30 seconds for it to capture the essence of something that took me several days to put together. 

But it just doesn’t feel the same. And maybe that’s the point.

Maybe we shouldn’t quit, after all.