Categories
Updates

Will ChatGPT (AI) replace Content Writers? In some cases, it already is

Back when I was starting out as a freelance writer, oh about 15 years ago, I was able make a modest income and sustain myself off a couple of writing gigs a week.

I had managed to land a gig with a handful of sites, with topics ranging from technology, geek culture, video games, food, health and wellness, and whatnot. I wrote several blogs a day at a fixed rate and collected enough income at the end of the month to pay for my apartment, unities, groceries, and expenses from university. I was by no means well-to-do, but hardly in dire financial straits.

While I haven’t been a freelance content writer for some time, I can say that for the possibility of one day going back down that route – or for people looking to get started in that manner – the chances of successfully being able to do so are becoming increasingly slim.

That’s not to say that the web is suddenly in less need of content than it was before, but rather that the ways in which sites go about creating it.

I’m talking, of course, about ChatGPT.

How did ChatGPT change the way we write?

It would be an understatement to say that ChatGPT has changed the way people write content in the digital age.

For the longest time, the most advanced software assistants were the squiggly red line in Word that called us out anytime it spotted a typo (real or imagined). Then came things like Grammarly for more advanced spell-checking and some basic auto-complete functions.

ChatGPT, for those of you not overly familiar with it, is essentially a chatbot crossed with an auto-complete function and then given a massive, impossibly large, dose of steroids. As a Large Language Model (LLM) it’s a bot that’s been trained to predict and complete text based on mountains of information that’s previously been fed into (I.E. most of the internet).

So, while on the surface, it is basically just a tool that figures out which word should come after the preceding word, it’s so good at doing so that it can generate text for you about virtually any subject, or answers for every question.

It’s important to note, however, that ChatGPT doesn’t inherently know things. If you ask it who the first Prime Minister of Canada was, it will likely correctly identify John A. MacDonald not because it knew that was the answer, but logically the words that should follow that question happen to be the correct answer.

In any event, getting into how LLMs and ChatGPT in particular work is a different topic entirely. What does matter is that these capabilities make the tool exceedingly strong when it comes to producing fairly mundane and common sorts of content. Sort of like the content you might read at any given moment on the web – like a blog.

How does blogging compare between ChatGPT and a writer?

Blogging is fundamentally the art of driving traffic organically to a website. It’s one of the most powerful ways to get a website to rank and to attract new visitors.

It’s also something of a numbers game. Sites that produce more content generally receive more traffic than those with a scarcer posting schedule. Also, sites that produce solid longform content also tend to reap traffic benefits for months to come.

The main issue holding back every website from being a blogging superpower (apart from talent acquisition) has largely been time.

ChatGPT is the tool that’s now disrupting this pain point. For better and worse.

Let’s do a comparison.

Calculate how much time it took A) your writer to produce that 450-word blog and then B) ChatGPT.

You’ll likely see that your writer can produce this piece in something typically ranging from 30-60 minutes while ChatGPT can complete its task in under 30 seconds.

As such, it will be the bot that takes the prize in every single case where speed is a factor.

In other words, if quantity is your sole consideration, you can produce more content with an AI tool than a human by a factor of several hundred.

However, we all know that blogging isn’t just about having the most content out there. For a blog to succeed it must also draw in readers, as well as retain them. Quality content, in this regard, will almost certainly trump quantity in the long run. After all, if your content is crap, readers will bounce and find what they are looking for elsewhere.

Fortunately, in the quality game, this is where human writers still have the edge. If you go back to our exercise and review the blog produced by our writer and that from ChatGPT, the human produced blog (assuming your writer knows their stuff) will almost always be much more readable, optimized, and likely to connect with your audience.

This becomes exceedingly true the longer your content requirements are. ChatGPT is strongest in short bursts but struggles to fill in longer pieces with meaningful text.

For instance, if you wish to create a standalone piece of evergreen content that’s 2,000 words, ChatGPT will get it for you in a minute or two, but the value it has for continually attracting visitors will more than likely be minimal compared to the human output.

A problem, however, is whether the people paying for content and using it to drive their traffic care about this distinction between quality for quantity.

ChatGPT content is already flooding the internet. Yup.

You don’t have to look very far to find AI created content popping up everything.

Websites such as Quora, which used to be a modest questions and answers site where neophytes could ask all manner of questions into obscure topics and the experts could respond, has exceedingly become a swill of nearly identical AI answers with the quality ones fighting for breath.

Affiliate sites, which I would guess already make up a significant portion of the internet, are seemingly becoming more common too, many of which are just using cheap, quickly regurgitated content across multiple domains.

There are also YouTube channels and podcasts of original kids’ stories that are AI produced, and likely with the help of LLMs providing the scripts and written content.

In some ways, it’s looking a lot like site owners are all onboard for AI content, and not the least bit worried about the grey sludge it’s creating across the web.

Are there no alternatives? Will it always be bot versus human?

The best answer I can give is no.

If you’re looking for both quality and quantity, there’s no need, or even reason, to force human and AI writers into silos.

Remember: at the end of the day, ChatGPT is a tool – not a solution. People who are using it to 100% generate their blogs, articles, and other written content are not using this tool the way it should be used.

To make the best of both worlds, and likely find a meaningful way forward, writers need to be not only familiar with tools like ChatGPT as competitors but also how they can use these tools to elevate themselves from competition.

The two ways writers can really benefit from this tool are 1) time to delivery and 2) quality of post.

In terms of delivery speed, we already know ChatGPT is quicker, but it’s not only quicker at writer. It can also do research, ideation, and outlining.

Using ChatGPT to help generate ideas for a client or put together a rough outline of a post in seconds can shave tons of time off each piece put together by our writer.

As such, while a blogger who delivers everything by hand, from initial idea, research, outline, and completed product, might take a fixed amount of time to complete the piece, by working a tool like ChatGPT into their routine they can not only increase their delivery time, but maybe even the quality of their output.

Ultimately, it comes down to knowing the right prompts and how to use the answers given by ChatGPT. But in the right hands, a talented writer can use the tool to help clear recurring areas of blockage in the usual workflow of article creation and complete their tasks more efficiently.

Now, the reason I’m bringing this up is not to suggest that speed is everything in a writer, but for freelancers who are often paid per completed article, being able to churn out more pieces in a given day can be the difference between paying rent or not that month.

For the second point about quality, it also ties to our time to deliver. If we as writers need to spend less time on the grunt work, it gives us more time to fine tune. With even just 15 minutes to spare from a task being handled by ChatGPT, we can use that to extend the length of our articles, add more details, or even further optimize the piece.

Why stop there? You can also use the Browse with Bing function of ChatGPT to look for sources in the background while you work.

When we start seeing what these tools can do for us and help improve our output, ideally, we’ll be in a place where we not only have a competitive edge with our delivery times, but also in terms of the quality of the articles.

And so, while we have unfortunately seen that plenty of sites have no qualms about low quality straight from ChatGPT rips, those that want something that’s a cut above will sooner, rather than later, be looking for the writers who can leverage these tools and offer more than those who see them as low-quality gimmicks.

It’s not necessarily the world I wanted to see, or even expected back when I was a full-time content writer, but like with every major technological change, there’s no turning back the clock. We can’t put ChatGPT or LLMs back into their box any more than we can the modern Office suite and word processor.

What does the future look like for writers in the ChatGPT era?

Humans, when given the time, resources, and proper assignments, at the present stage in history will always be able to write a better blog post, article, short story, product review, what have you, than anything produced entirely by the AI. The main drawback, however, is time – and in the cases where a business is hiring a writer – the money as well.

As we’ve pretty much seen throughout history, paying people to complete tasks typically costs more and takes longer than having a machine automate the process. Whether it’s an old-fashion cobbler working on a single pair of boots compared to a factory churning out dozens by the minute, online writers are now facing the same challenges as craftspeople throughout history have.

The craft associated with content won’t entirely die out. While the majority of people buy their clothes and attire from retailers at the mall, outlet, or online, there are still master artisans tailoring clothes to order. While IKEA has changed the way furniture is made and distributed, there are still carpenters able to build you a table based on specifications.

There are, however, far fewer of these artisans – and those that do survive economically are likely masters of their particular niche. In this manner, our master storytellers churning out the top selling fiction we can’t stop reading will likely carry on for the near future without much disruption. However, it’s all the other writing that we typically encounter on a day-to-day basis (and largely online) which is going to exceedingly be AI-written.

Web content, I suspect, is going to follow a similar route. The websites that produce the most content will inevitably be those that put AI into their production workflows. The caveat being, that if they rely too much on AI and not enough on human intervention then they certainly won’t be the sites with the best content out there.

The future, in this case, will be a hybrid, a synthesis, however comfortable or uncomfortable that makes us.

Categories
Updates

“The Roller Coaster” – Surreal Short Fiction, out via Litro

Photo by Augustine Wong.

My surrealist short fiction piece – “The Roller Coaster” just came out today with Litro UK, in their lunch break fiction section.

It’s a small piece that’s pretty indebted to the style and trappings of Haruki Murakami’s short fiction, which I had read quite a bit of last year when I first wrote it.

In essence, a young man living in Montreal meets someone at university and they become very close. However, their relationship and his own perception of nothing short of reality come into question one afternoon when they visit LaRonde together.

Categories
Updates

Books I Read in 2023

I finally tallied up my reading list. In 2023, I read 33 books. Most for the first time, a couple for the second, third, or possibly fourth time.

The list is:

James SA Corey – Tiamat’s Wrath
James SA Corey – Leviathan Falls
Jane Harper – The Lost Man
Something is Killing the Children, Volumes 1,2,3
Keigo Higashino – Malice
Keigo Higashino – A Midsummer’s Equation
Keigo Higashino – Newcomer
Keigo Higashino – Silent Parade
Arimasa Osawa – Shinjuku Shark
Stephen King – The Outsider
Daniel Keyes – Flowers for Algernon
Michael Crichton – Eaters of the Dead
Richard Stark – Break Out
Richard Stark – Flash Fire
Richard Stark – Dirty Money
Richard Stark – The Green Eagle Score
Krishnadev Calamur – Murder in Mumbai
Nicholas Carr – The Shallows: What the Internet is Doing to Our Brain
William Gibson – Neuromancer
Jussi Adler Olsen – The Scarred Woman
Jussi Adler Olsen – Victim 2117
Jo Nesbo – The Bat
Jo Nesbo – Redbreast
Jo Nesbo – Nemesis
Jo Nesbo – Cockroaches
Jo Nesbo – The Devil’s Star
Jo Nesbo – The Redeemer
Jo Nesbo – The Snowman
Jo Nesbo – The Leopard
Robert Sellers – Hellraisers
Kanae Minato – Confessions
Ken Baumann – Earthbound
Sydney Sheldon – If Tomorrow Comes

I started they year out by completing the Expanse series. I hadn’t reach much sci-fi, borderline space opera for a while, and it made for a pretty gripping 7 book series. Though, whether I will go back and reread them any time soon, considering I’ve seen the television show twice over, is unfortunately another matter (even if the books, as usual, are better).

After having only previously read “The Snowman” by Jo Nesbo and found it sillier than the average Nordic Noir, I’m glad I gave the author another chance. I read a bunch of the Harry Hole books this year, mostly out of order, and was pleasantly surprised that the quality was not only better, but so too was The Snowman in the context of being the sixth or seventh in the series. I had planned to read more, but after losing my kindle and a couple of half read books on it (including another Jo Nesbo), I haven’t had the stomach yet to buy another device and pick up where I left off.

Sydney Sheldon probably wins the award for writing the cheesiest book I read all year. Despite that, it was somehow remarkably readable and a guilty page turner, kind of the way other mass market paperback thrillers tend to be.

The Richard Starks were ones I had previously reread (and at least 2 on trains this year). Reminded me for a time this summer that I need to pick up more. However, he’s one of those authors I always scour for at used book shops and he hasn’t turned up at the past couple I checked.

Shinjuku Shark by Arimasa Osawa was a delightfully gritty hard detective noir that I wasn’t expecting to like so much. Definitely a Raymond Chandler type set in 1980s or 90s Japan. I also found out the book is fairly rare, and its not too probable that I will find the following ones in English either. There was maybe one print a while back, and I haven’t seen them listed anywhere except for exorbitant prices on Amazon.

Lastly, Neuromancer. This was my third or fourth time reading it. It’s one of those books that changes each time I do, especially the protagonists. Where once upon a time Cage and Molly were these larger than life adults, living strange lives that were years ahead of me, I’m now substantially older than both of them. I couldn’t help but see a naivety and an innocence to them now, despite both of them being hardened criminals who had somehow seen it all in their early to mid 20s. It’s nice to think people could be so young and so pessimistic.