Should brands trust AI tools to write their content?

  • Categories:

    Content Marketing, Industry Trends, Search

  • Date:

    August 15, 2023

Should brands trust AI tools to write their content?

Content Marketing Industry Trends Search

While artificial intelligence isn’t new, 2023 is, without question, the year of AI. Alongside this surge of excitement, many marketers are experiencing pressure to jump on the bandwagon lest they get left behind. But are we in for the long haul, or is AI just having a (powerful) moment?

In an age where AI and large language models have made significant strides in a short amount of time, I’ve thought a lot about one crucial question: Should brands trust AI writing tools to write their content? Is “writing” even the right way to describe what these tools do?

The fact is, there are no simple answers to the matter of if, when and how to leverage AI technology for writing. It’s a complex topic involving work quality, productivity, data security and more — too much for one little blog post to tackle.

For now, then, let’s take a step back and look at the fundamental differences, benefits and consequences of using AI tools for written content in 2023, along with how humans and machines can (and should) work together.

What are AI content writing tools, and how do they work?

First, let’s take a moment to recognize that every piece of content likely has at least a little AI in it. Spell-checkers made a splash on PCs in the early 1980s, before yours truly learned to read. Autocorrect comes standard on Android and Apple smartphones (though this feature is fallible, as anyone who has ever uttered the words, “Damn you, autocorrect!” knows all too well). Writing assistant Grammarly is so prolific, its co-founders became billionaires in 2021.

But the technology making headlines in 2023 does more than proofread human writers’ work. AI content writing tools such as ChatGPT are software applications that use AI to help people write content. By using a combination of natural language processing and machine learning, these tools can generate long-form content such as blog posts or landing page copy. They can also be prompted to summarize long text (e.g., by providing bullet points for a resume).

Most of these tools apply NLP to understand and process user prompts. Without NLP, AI content writing tools could not recognize the relationships between words and comprehend what the user is trying to say. Furthermore, machine learning algorithms can train AI content writing tools to generate text that is similar to the data used to train the computer.

In short, AI writing tools employ NLP to understand the user’s prompt and ML to provide the most appropriate output to the user. But the how is far more complicated, as popular AI writing tools rely on a number of factors to produce accurate, informative and relevant text.

Benefits of Using AI Tools for Content Creation

Overcoming writer’s block

With the right prompt engineering, AI writing tools can become a welcome salve for writer’s block by providing a faster approach to short-form items such as headlines, subject lines and preheaders. Instead of staring at a blinking cursor and a blank page, writers can take the fast lane out of writer’s block purgatory by asking ChatGPT, for example, to come up with three options for a headline or a list of tips for a blog post based on notes they provide. Then, they can edit from there.

Improving productivity

Scaling some of the repetitive tasks that were previously manual by utilizing AI content writing tools can allow humans to spend more time building a strategy and creating core content assets.

But successful marketing takes more than speedy output, and brands shouldn’t use time savings as an excuse for lower-quality content. At Wray Ward, we see AI writing platforms as tools that will allow our writers to create more and better content by using AI to tackle the repetitive tasks — while humans do the deep thinking that results in impactful content.

Brands should look at generative AI not as a means of being more creative, but instead as a means of shifting existing work processes so that professional writers have more time to be creative. Effective use of AI tools can give us more creative freedom simply by giving us more time to think.

Creating derivative content

Often, after our content writers develop a long-form, high-value piece of content, they’ll adapt this content for multiple assets across various channels. It’s an efficient way to operate, but it’s also smart because it helps forge natural connections between the assets and builds a journey for the target audience to follow.

Creating derivative content can be time-consuming and repetitive, but it’s just as important as the core content asset. AI tools can be a great solution for quickly generating shorter versions of blog posts, emails, organic social posts and more.

However, these derivative assets will only be as strong as the original content, and all prompts aren’t created equal. Knowing how to effectively prompt the AI tool is a skill, just as knowing how to use any other kind of technology is a skill, and outputs should always be considered first drafts.

Repurposing existing content

Repurposing existing content may be one of the most rewarding but also most overlooked tasks, partly because it — surprise — takes time. Almost every site contains a landing page with shallow content or a blog post or news article with an outdated statistic or too-long title tag. Using generative AI tools may be an efficient approach to revising existing content while extending its shelf life and value. Using solid prompt engineering to repurpose existing content can also allow brands to target more specific audiences or search engine results page elements.

For example, we used the LinkReader plug-in on GPT-4 (now available with ChatGPT Plus) to scan a URL from an older blog article on our site. After giving the tool context regarding the target keyword and subsidiary topics we wanted to focus on, we used it to revise the content and answer a specific question more clearly.

We also tested GPT-4 by asking for feedback about how well a segment of a published blog post answers a specific question. While the tool was not able to write content that could be used directly in the article, it gave some useful starting points on how to improve the existing content for specific keywords or People Also Asked questions.

What are the risks of writing content with too much reliance on AI?

Brands that rely too much on AI run the risk of publishing unoriginal content.

Storytelling is a fundamental part of being human. In fact, a 2021 study found that the hippocampus, a complex structure buried deep in the brain, constructs stories by connecting seemingly unrelated facts and events into coherent narratives. Wray Ward’s clients rely on our people for deep-brain magic: creative, compelling content that drives meaningful action.

On the other hand, AI writing tools rely on existing data and works by other AI writing tools or human writers to generate content. This means that using AI writing tools increases the probability that the content output will lack originality or uniqueness. In fact, brands that rely too much on AI risk becoming indistinguishable from their competitors. Not only that, but AI writing tools can easily generate plagiarized content from other sources — and the plagiarized content may not be easy to catch until it’s too late. Talk about a legal can of worms.

This doesn’t mean that AI writing tools are ineffective during the brainstorming and experimentation stages. But if we shift too much of the responsibility of storytelling from humans to AI, we risk commoditizing our brand.

AI writing tools may struggle to understand your target audience.

Brands that understand their audience and can address their unique needs will always win over the competition. On the other hand, nothing will resonate unless it springs from an understanding of the target audience’s mindsets, motivations, media habits and media consumption. The audience is key to everything we do, and it isn’t as simple as baseline demographics: Marketers have to synthesize a wealth of sources in order to know how to get the audience’s attention and drive meaningful engagement.

HubSpot claims that its content assistant, powered by OpenAI and released in March 2023, can help marketers “not only streamline busywork, but elevate the quality of the work they love.” Yet as the company said in a 2015 blog post, “it’s much harder to create compelling content marketing if you don’t know why it would be compelling to your audience in the first place.” That’s where a complete understanding of the brand and its audience(s) becomes crucial. Such knowledge is not readily available and is often proprietary.

To test the ability of the HubSpot content assistant to cover a common topic, I prompted it to “write an engaging email about a guide that includes tips for creating blog content that gets results.” Here’s what I got:

  • Hero Article Title

  • Hero article description here

  • Dear [Name],

  • Are you tired of spending hours crafting blog content that seems to fall flat? Do you want to learn the secrets to creating blog posts that actually get results? Look no further! Our new guide is here to help you become a blogging superstar. Packed with expert tips and strategies, this guide will teach you how to craft compelling content that engages your readers and drives them to take action. From headline writing to keyword optimization, we cover it all. Don’t miss out on this opportunity to take your blog to the next level. Read More to get your hands on our ultimate guide to creating blog content that gets results.

  • Best regards,

  • [Your Name]

This draft gave me a few thought-starters. However, armed only with my basic prompt, the AI tool skipped the heavy lifting of analyzing and developing a complete understanding of my customer, their challenge and what made my brand uniquely positioned to provide a solution.

AI writing tools cannot mimic your unique brand voice.

HubSpot’s content assistant can’t simulate the knowledge of an insights team, but can it capture your brand’s voice? The AI tool can generate assets from prompts, and it can also update those assets. So, I asked the assistant to improve on the email it had just written for me. I highlighted “help you become a blogging superstar” and brought up the options to change the tone.

  • Friendly

  • Professional

  • Witty

  • Heartfelt

  • Educational

I selected “friendly,” and — voila! — that piece of text became “help you become a blogging sensation.”

That’s a stretch. It’s a little like using a thesaurus without understanding who you’re talking to or what you want them to do. Consider, for example, that lists “affectionate” as a synonym for “friendly.” But the nuances of human language are such that those two words could send a totally different message depending on the audience and the context.

It was just one exercise, but I walked away thinking this: While helpful in experimentation and to jog the brain when it’s suffering from writer’s block, don’t mistake HubSpot’s content assistant for a do-it-all wunderkind. We will always need creative and strategic thinkers to conceive and protect brand voice as well as emotion, timeliness and novel perspectives.

(In fairness to the folks at HubSpot, they acknowledged that they are NOT relying heavily on AI to generate content. Instead, they edit the output for voice and other considerations and conduct follow-up research to ensure accuracy. They also acknowledge that AI tools are more useful for help with basic tasks.)

Privacy should be a big consideration for brands using AI.

AI content writing tools rely on users to feed their ML algorithms and constantly improve and strengthen their outputs. Some, such as OpenAI, assume ownership of the inputs they receive. While this may help AI tools become more effective over time, brands should be cautious about the information shared within their prompts, as this is used to train the tool — and sensitive information may be shared with other users.

AI writing tools are inherently biased.

Going back to ML and how computers are trained, AI content writing tools depend on the data they are trained on. This can create biases in the outputs of AI writing tools — biases that experienced, skilled human writers are able to detect and avoid.

Will AI replace human content writers?

The short answer? No. We live in a world of misinformation, and because AI tools are not consistently factual, there will continue to be an increased need for humans to draw on personal experiences and knowledge to fact-check content that is published for their brand.

Not only that, but a people-first approach to content that demonstrates E-E-A-T (experience, expertise, authority and trustworthiness) is becoming more important than ever and should be a key focus for content marketers. AI-generated content is less likely to succeed at E-E-A-T and SEO, because search engines prioritize content based on a level of experience and expertise that generative AI simply can’t replicate.

At the end of the day, the best communicators understand the nuances of human psychology better than any machine, because they live that human experience every day.

Finding the Right Balance of Human–AI Collaboration for Content Creation

There’s no denying that AI is here to stay, and it would be foolish to completely dismiss the aid of AI tools for writing content. Instead, we must make sure we select the right AI tools and use them appropriately, in a way that amplifies the skills and ideas we already have. We must also be willing to recognize the shortcomings of AI tools as well as their virtues.

Since the Stone Age, humans have used tools to make their work more efficient. And in this Age of AI, we have an enormous responsibility to do it thoughtfully.

Explore more articles from Wray Ward.