Hardly a week goes by without seeing one of those infographics or social posts describing why ChatGPT (or other generative AI) is coming for certain jobs—including content writers, copywriters, and creators.
So you may be asking: should we replace our human content writers with generative AI?
As tempting as it may seem, please don’t make this mistake. Unless, of course, you want your content quality—and performance—to suffer.
Let’s start with the basics of generative AI
Since you’re reading this article, I’m going to assume that you’re at least somewhat familiar with generative AI and ChatGPT in particular—if so, feel free to skip to the next section.
But here are some insights into how generative AI works that you need to understand for this article to make sense:
- As the name implies, generative AI ingests data, analyzes trends and commonalities, and generates similar outputs to those data.
- Generative AI, like any machine learning algorithm, must be trained on specific datasets to work—and its output is limited to whatever data it ingests (for example, ChatGPT 4 is limited to data before September 2021)
- There is no guarantee of the quality or accuracy of generative AI’s outputs—it all depends on training data quality and whether false, misleading, or otherwise erroneous outputs have been corrected (which, as we’ll see below, requires human input)
3 dangerous misconceptions about generative AI & content writing
Now that we have those basics down, here are three common, but dangerous, misconceptions about generative AI and how it relates to content writing.
1. Generative AI can generate high-quality content
Let’s start by ripping the band-aid off: generative AI cannot write high-quality content.
In a recent Guardian article, a copywriter claimed ChatGPT could generate an article that, with a bit of punching up, would be identical in quality to one of his own.
If that’s truly the case, then he’s wasting people’s money. Here’s why:
- Generative AI models are not trained on a particular business’s customers, pain points, marketing strategies, messaging and positioning, etc. (unless you have a highly sophisticated, proprietary model—in which case, let’s talk!)
- Generative AI is excellent at aggregating and surfacing information, but not at making a persuasive argument—and it doesn’t handle nuance all that well
- Generative AI is hallucination-prone, meaning that you can’t automatically trust the accuracy of its outputs
If you want to create content that will persuade an audience to buy and anticipate their objections, you shouldn’t rely on generative AI.
The only content writers generative AI will replace are the $0.05/word copymills. But if you care about creating valuable content that educates and informs your audience, you’re not going to work with them anyway. 😉
2. Generative AI is automatically a value-add
Whether we’re talking about predictive or generative, deep learning or neural networks, or any AI jargon, many marketers mistakenly treat “AI” as a differentiator. Just look at these article titles:
- The Always Up-to-date List of the 50 Best AI Tools in 2023
- 17 SaaS AI Tools That Are on Fire in 2023
- 20 AI Tools To Run A Productive SaaS Business in 2023
While they’re probably optimized for AI-specific keywords, they miss the mark in terms of persuasive marketing. Why? Because they promote features over benefits.
Few buyers will care that a product uses AI as part of its technological architecture. Rather, they want to know: “what will AI do for me?”
Knowing a product is “AI-powered” doesn’t mean anything. It’s a technical spec, not a benefit. They’re really only relevant if you’re selling to highly educated, technical buyers.
So why is everyone saying “AI” this and “AI” that? Simple: it’s a fad. And brands want to signal that they’re in on the fad—it probably gets them in the door with investors and some PR coverage.
But while AI isn’t itself a value-add, like any other technology or tool, it can offer certain benefits:
- Saving time by generating more content faster. But you still need significant human input to edit, validate, and improve the content.
- Saving money by reducing personnel overhead. The problem here, of course, is that poor quality content hurts a brand’s reputation and underperforms in generating demand. In other words, the money you save could be displaced by lost revenue.
- Closing the loop on a particular writer’s blind spots. This can, in my view, be a huge benefit—with AI, I can bring myself up to speed on certain topics faster than before. However, remember that AI is sometimes erroneous, so while it can augment the human writer’s knowledge base, it can’t replace it.
We could keep going down the list. The fact is, AI itself isn’t an automatic value-add. It’s what you do with it that matters. This is the same for your products as it is for content writers.
3. It’s “either-or”: generative AI must replace all human creative input
Most people involved in building AI-based products don’t make this mistake. Unfortunately, too many people posting about AI online aren’t involved in building AI products, so this myth tends to perpetuate.
AI, by its nature, can’t work without some human input. This happens in three dimensions:
- Building & training the model. Humans have to first build the AI, then continually adjust it based on the quality of its outputs. Determining what “good” looks like requires human input.
- Preparing the data. AI models are trained to optimize their outputs based on inputted data. The quality of that data, then, determines the quality of the model. Humans are indispensable in driving that data quality.
- Applying the model’s outputs. Once the AI generates an output, a human being must determine what the next step should be.
Although generative AI can replace a number of monotonous tasks, it cannot replace the human being outright.
For content writing specifically, there will always need to be a human “editor” to review AI output, determine its quality, and adjust in real time. What’s more, the quality of a generative AI output often depends on the prompt you use.
This, of course, requires human input and intuition, and an already mapped-out strategic goal to implement.
Why you shouldn’t rely on generative AI for your content writing
So as a business or GTM leader, what does this mean for you?
Fundamentally, you need to be clear-eyed about the risks of using generative AI to write marketing content.
Because if you do, the quality of your content—and, thus, your marketing performance—will suffer. Here are four reasons why this is the case.
Your content won’t stand out.
Generative AI is trained to ingest existing data and generate similar outputs. Humans, on the other hand, have a unique knack for creativity, approaching problems and attacking them in different ways.
So if you rely on AI-generated content, you’re basically regurgitating information and points of view already out there. Why, then, should someone read your content when they can get the original somewhere else?
Your content won’t resonate with readers.
I’ve come up with a sniff test for determining whether an article is AI-generated. If it’s dry, boring, and verbose, then I check to see if AI wrote it (and I’m usually right).
I’ve never seen an AI generate content that actually connects and resonates with a human reader. 99% of the time, it’s “just the facts, ma’am.” No voice, no point of view, and no spirit.
If you had a salesperson that sold like that, you’d fire them instantly. So why would you let generative AI sell to your prospects—at least without a human writer to edit, adjust, and enhance its output?
Your content won’t align with your brand voice.
Whether you’re B2B or B2C, you’re selling to human beings. And the more humans are connected with your brand, the more they’ll spend—more than half of consumers (57%) will increase their spending with that brand and 76% will buy from them over a competitor.
Generative AI, however, can’t create content with a distinct voice that connects with human readers—because generative AI can’t create content with distinct anything.
If you use AI to create your content, you won’t foster those connections that are key to building brand loyalty and customer retention.
Your content won’t persuade your audience to buy.
Generative AI cannot answer the question: would a human reader buy this argument? Human intuition is necessary to gauge whether content is going to persuade the reader to take further action and, ultimately, buy your product or service.
So while, yes, generative AI is a powerful tool to have in the bag, it doesn’t think on this human level. As such, you’ll need a human writer to review the content to make sure it will actually drive the reader to learn more and, eventually, buy your product.
Final thoughts on generative AI and content writing
Generative AI is to the word processor what the word processor was to the electric typewriter. In other words, it’s a tool.
Like any other tool, it can be used for good or ill.
Personally, I find AI helpful in ideating on content topics and speeding up base-level research in new subject matters. But I never publish a single word of a generative AI output. Rather, I use the output as one source of information among many—and use all of them to bolster the argument and create something unique.
When I spoke with Jordan Ogren about this recently on A Podcast About Marketing, he referred to AI as a “second brain.” I think that’s pretty close to the truth—except for the fact that, unlike the human brain, it can’t function on its own in the creative realm.
So I’m not worried about AI coming for my job. And, as you look to ramp up your content marketing endeavors, you shouldn’t rely on it solely to generate your content.
Rather, find writers and content marketers who know how to manage the best of both worlds. In my view, that’s the future of content marketing.
Since human content writers are still table stakes, you need one on your team. Let’s set up a time to discuss how our team can help you.