Will AI Replace Content Writers?

While we can never underestimate the power of machines, AI is yet to replicate the nuances of tone, niche language, subjectivity, and authenticity through personal experiences.

For these reasons, it’s hard to see a future where content writing will be fully automated.

However, as writers, we must understand the landscape of the market.

Bloggers will be left in the dust if competitors find ways to leverage AI to produce higher volumes of quality content.

Mediocre freelance writers will miss out on client work as robots become more capable.

In this post, I’ve made it super easy for you to understand the impact of AI on content writing so you can stay on top of your game.

How AI writers work

AI, short for artificial intelligence, is the simulation of human intelligence performed by machines.

In the context of writing, the success of AI is measured by how well robots can replicate human writing.

The company pioneering AI writing is OpenAI, co-founded by Elon Musk and run by Sam Altman, former president of Y-Combinator.

OpenAI has developed a protocol called GPT-3 which stands for version 3 of the Generative Pre-trained Transformer. Stay with me here.

GPT-3 is a language prediction framework that learns text structures through collected data and then tries to predict word formulations based on the information.

GPT-3 crawls the internet for obscene volumes of text through publicly available datasets. We’re talking about over 570GB of text information collected!

What separates GPT-3 from its predecessors is the rigorous training the AI has undertaken to improve its accuracy. OpenAI has invested over $4.6 billion (and counting) in GTP-3 pre-training to understand human language through machine learning.

In summary, GPT-3 is predictive text on steroids 🚀

Then we have software companies like Jasper that use GPT-3 to build AI writing assistants.

Below, we’ll look at what an assistant like Jasper can do by leveraging the GPT-3 framework.

What type of writing can AI do?

As GPT-3 is an ever-growing neural network of text data, the possibilities of these AI writing assistants are seemingly endless.

AI can perform just about any writing task from headlines, sales copy, product descriptions, ideation, outlines, articles, and even books.

The only example of AI writing being a stretch is fiction, which relies heavily on human imagination and experiences. Wait, hold that thought. This is an excellent segue to what AI struggles with in relation to content writing.

The shortcomings of AI writing assistants

Modern internet writing often involves extensive keyword research, outlining, drafting, editing, formatting, and optimizing. From this perspective, AI is an invaluable resource for publishing content that ranks and converts.

As someone who makes a living splitting time between building authority blogs and client work, I felt both excited and threatened by AI.

I thought AI writing would provide infinite scale.

The proposition is compelling. Pay $59 per month and get unlimited written words. My imagination went wild with ideas.

For example, I could do keyword research on the Kindle store to find book ideas, then use the AI to create book outlines and even write the drafts. With this process, I could self-publish dozens of books in a year, starting a nice residual income business.

Additionally, it takes anywhere between 5 and 20 hours to publish a single article. Surely AI could help me to at least half my production time.

Conversely, as a freelance writer, I thought, shit, most entrepreneurs and project managers probably feel the same as me—meaning a good chunk of the freelance market will be replaced by AI writing tools.

But after testing apps like Jasper, admittedly, I was underwhelmed.

I know it’s early days, and the AI will only get more “human-like” as it processes more data—but still, I have my doubts.

Update: I tried Jasper again, pushed through, and have found great success. Not for highly technical client work, but certainly for B2C lifestyle content on my niche websites. More on that later.

Below are four areas in which humans have the advantage over robots (for now 🤖).

1. AI doesn’t understand community language and cultural references

The internet, heck, the world is built on the foundation of niches. People identify and participate in different “clicks,” and each community has its unique language.

For example, I love basketball, Naruto (anime), blogging, and Magic The Gathering. I’m also vegan.

All these niches have their own set of nuanced language and references that only that community would understand.

Beyond just understanding how to compile and construct plagiarism-free sentences, good writers also need to put themselves in the shoes of their audience and speak their language.

How you write to an audience of stay-at-home moms differs from how you’d write to real estate agents.

I mean, can AI really understand the frustrating ending to Game of Thrones? Here’s Jasper’s opinion on GoT with no context:

In the final episode of Game of Thrones, many fans were left frustrated by the way the show ended. Some characters who had been in the show for many seasons met their fates in a way that was unexpected and unsatisfying. Daenerys Targaryen, who had been one of the show’s main characters since season one, was killed by her nephew Jon Snow, after he realized that she was planning to genocide all of the people in Westeros. This anticlimactic ending was frustrating for many fans, who felt that the show had let them down.

Technically, it’s okay, but it’s too objective and boring. The AI fails to capture the emotion of the show’s audience.

Also, with social media, we have new cultural trends every day. Musicians are literally creating music to trend on TikTok.

How would AI keep up with quick nuanced shifts in language and culture? I’m willing to bet it will struggle.

2. AI doesn’t have any personality

AI writers fair pretty well in processing and constructing sentences on a chosen topic. However, readers demand more than just information.

We want to be engaged as we read. Being able to communicate wit, charisma, and emotion is not programmed into the GPT-3 protocol.

In saying that, I’ve been impressed with Jasper’s ability to incorporate tone into its output. For example, in the long-form editor, you can command the AI to write in Donald Trump’s voice, which is amusing.

I’ve found a lot of success setting the tone to “relatable” as Jasper attempts to write in the first person with a hint of empathy.

The internet is shrinking our attention spans, and writing with personality could be a big differentiator moving forward.

3. AI doesn’t write in the first person

It took me five minutes to find an SEO writing job on Upwork, where the client specifically wanted content written in the first person.

Writing in the third person has its place. However, it can also hinder the ability to form a personal connection with readers.

Writing in the first person allows you to share personal stories and experiences relevant to the reader. Humans are nosy, and we love to look at how others think and live.

As AI isn’t a person, it can’t draw upon its own experiences to tell a relatable story.

Update: I’ve since discovered you can get Jasper to write in the first person, especially if you write a few sentences in that style. However, you need to constantly feed it information for the AI to output accurate examples—after all, these are your experiences.

4. AI doesn’t have an opinion

Could you imagine if all content was produced without an opinion? Talk about a boring read!

The ability to form an opinion and be subjective polarizes audiences and increases engagement—sometimes to our detriment.

Good writers switch gears between subjective and objective angles to keep the reader on their toes.

Sometimes you must draw a line in the sand to build super fans while simultaneously attracting enemies. You’re either in, or you’re out. I can’t see a path where AI would be able to pull off such a feat.

How can AI technology help us write?

While there are some fundamental areas where AI is lagging, there’s still a massive opportunity to leverage tech to create superior SEO content.

For example, you can use Jasper to generate topic ideas and even write some of your content. SEO tools like Frase or Surfer to optimize your writing for search. And Grammarly to professionally edit your content.

Keyword researchAHrefs, SEMRush, KWFinder
BriefingFrase, MarketMuse, Surfer SEO
WritingConversion.ai, Copy.ai
SEO optimizationFrase, MarketMuse, Surfer SEO
EditingGrammarly, Hemmingway

It’s interesting, though, because most of these services position themselves as assistants. We enter murky water when we try to fully automate the content writing process with tech.

AI writers should be used sentence by sentence in collaboration with a human writer to push through blockages while keeping those intangible human elements.

SEO optimization tools are great to ensure you’re not too far off the mark. However, the more writers use these tools, the more similar each ranking piece will be. So be careful not to over-optimize and potentially lose your differentiation.

Then with Grammarly, while incredibly powerful, it’s still not 100% accurate (it’s close). You still need to review each suggestion carefully to ensure the AI doesn’t completely change the meaning of your writing.

In summary, AI is merely a tool, not a replacement for the craft of content writing. Work in collaboration, and you’ll be sure to improve the quality and performance of your writing.

FAQs about AI writers

Q: Can GPT-3 replace writers?

GPT-3 can technically produce any non-fiction writing. But, GPT-3 can’t, at least for now, can’t articulate nuanced niche-based language, authentic stories, opinion pieces, and creative fictional writing without extensive input.

Q: Can AI create a meaningful story?

AI is not a person. Thus, it can’t share stories based on the writer’s personal experiences. However, in theory, AI could curate and construct narratives based on other people’s experiences to deliver a meaningful summary.

Q: Can AI create original content

AI is constrained by data that is already publicly available. So, in essence, the AI communicates established information slightly uniquely.

AI can’t do original research like running surveys, interviewing people, conducting focus groups, or testing subjects like humans can.

Q: Will AI replace copywriters?

AI writing assistants are surprisingly good at producing concise sales copy for emails, digital ads, product descriptions, and landing pages. Yet, the effectiveness of AI copywriting depends on the quality of the instructions you give.

In other words, AI still needs direction to address target customers’ pain points and incorporate features and benefits. And with a ton of revenue at stake, you’d want to oversee the writing process.

Similar Posts