Write-up: Key takeaways from my academic research on generative AI in strategic communication

In spring of 2023, I worked on an academic research project on the topic of AI in strategic communication with the Department of Communication at the University of Colorado, Colorado Springs.?

I learned so much about how we can partner with machines to create smarter, more effective marketing and sales communication. However, I also discovered more potential pitfalls than I expected.

I’m personally blown away by the possibilities with generative AI (aka GenAI), but through my research, I have come to understand that the average user doesn’t usually have the whole picture with this tech.?

In this article, I’m going to share what I learned in my graduate studies — and my hope is by the end, you’ll have that whole picture.?

Powerful GenAI content creation use cases

Conquering the blank-page blues

When we’re creating content of any kind, many of us experience what I call the “blank-page blues.” It’s that paralyzed, sometimes anxious feeling that happens when you’re staring at a blank page.

If you struggle with this, GenAI tools like ChatGPT can be incredibly helpful for ideation.

AI can help you:

• Generate topic ideas.
• Create outlines or bullet points about your topic.
• Kickstart your thinking as you’re trying to organize your thoughts.
• Brainstorm headlines or subject lines.
• Choose appropriate keywords or social media tags.

Expanding your vocabulary and brainstorming analogies

I frequently use ChatGPT as a thesaurus to help me find more creative words and phrases. I give it very specific parameters, like “Please give me 10 other ways to say ‘creative’ that start with the letter C.” (For the record, it still gives me words that don’t start with C in the output, but I digress.)

It’s also great at brainstorming analogies. I’m a word nerd, but I struggle to come up with good analogies sometimes! And analogies are incredibly powerful for helping people understand new or complex concepts.

Testing ideas

One of my favorite ways to use GenAI tools is to test ideas. For example, if I have an idea for a workshop or class for writers, I give ChatGPT a prompt with my target audience, the objective of the course, and the high-level modules and objectives.

I ask the bot to help me flesh out the modules, then clean it up to create a “minimum viable product.” Finally, I run the full outline by a handful of carefully selected critics and ask them to poke holes in it. If they come up with too many problems, I don’t pursue the idea.?

This can save you tons of time when you’re planning campaigns or figuring out your next sales and marketing move.

The limitations and problems of GenAI

Machines can’t provide nuanced human narrative

Narratives that are rich in nuance and emotional resonance are distinctly — and perhaps exclusively — human.?

Algorithms, no matter how advanced, can’t truly grasp meaning. And without meaning, the core of a story evaporates. The text is just a collection of words without a soul, missing the narrative pulse that gives it life. Even the most enigmatic poetry carries a story, a heartbeat, when written by a human hand.

This is one of the biggest reasons GenAI output can sound like it’s been written by a bot, and why we must take the time to edit any text we get from tools like ChatGPT (you’re going to hear me say that again and again!).?

If content is going to resonate with your audience, it’s gotta have soul.

GenAI can’t replicate individual human experiences

In a world of content overload, we’re still eager to hear stories from people who have been in the trenches and come out with insight.

AI technology holds enormous potential, but the one thing it can't replicate is human experience. This is why unique, well-written thought leadership content still works so well, and probably always will.?

Models make small mistakes — as well as glaring errors

If you’ve used these tools before, you’ve probably noticed they often make small grammatical errors. I’ve spotted several syntax and sentence structure issues ChatGPT tends to repeat. The dangling modifiers alone are enough to drive me up the wall — and if you don’t make an effort to edit out these mistakes, your content will sound stiff at best and inaccurate at worst.

“Small grammatical errors” may seem inconsequential — but there’s a cost to mistakes like these. If it creates even a split second of friction for prospects and customers, it can hurt your bottom line.

Generative AI models make bigger and more devastating errors, sometimes, though, too. “Hallucination” — a confident answer from a language model that is not based in fact (aka, it’s a lie) — is a known issue. And you can’t fault the machine, because it’s not human. Generative AI doesn’t have ethics, or know fact from fiction; it’s a prediction engine. Tools like ChatGPT are purpose built to use their learned datasets to predict the next right words. Making sure those words are accurate and factual is the user’s job.

Using generative AI to write your content doesn’t actually save you time

In one of the more startling studies I found, it was uncovered that in a business setting, using ChatGPT in the writing process cut writing time in half … but it doubled the editing time.

So, using AI didn’t actually save users any time.?

That said, this may be a case for strong editors to use AI more often when creating content. If writing a first draft is a slog, but editing is a pleasure, using generative AI may be a benefit.

Guidelines for companies and organizations that want to use generative AI tools for content creation

Don’t use output as-is

There's an irresistible allure to the speed and efficiency of AI-driven content creation. But you should never take AI content, publish it word for word on your website, and declare the job done.

Treat AI output as a starting point in your content creation. Your role is to sculpt and refine this raw material to turn it into a compelling piece of communication.

Marketing and sales are all about engagement with your audience, so your goal should be to genuinely connect with the reader — not just put words on a page. Make sure your narrative answers the big question, “Why should I care about this?” and takes readers on a journey with a clear beginning, middle and end.

Continually refine your prompts to add specificity and nuance and get more sophisticated first drafts. Then put your drafts through rigorous fact-checking and editorial scrutiny before hitting publish.

Make sure your processes are in place before you introduce AI?

The buzz surrounding AI tools like Jasper, ChatGPT and Writer.ai is palpable, and many marketers are eagerly picturing a future when GenAI tools streamline and improve content production.?

But a word of caution is in order: Before introducing AI into your content ecosystem, make sure your foundational processes are rock solid.

For businesses that have seasoned copywriters and meticulous copy editors on their teams, integrating AI can be transformative. But companies that delegate AI-generated content to novice freelance writers, without a supervisory framework, are treading on thin ice.?

Without oversight, the result is content that's disjointed and subpar, at best.

Before integrating AI into your content creation process, establish robust processes that emphasize:

• Consistency: Use a uniform voice and style across all content.
• Narrative strength: Engage readers with compelling storytelling.
• Brand alignment: Use your content to reflect and reinforce your brand identity.
• Effective CTAs: Motivate your users to take the next step by using clear and enticing calls to action.
• Positive user experience: Make sure your content is relevant and accessible, and that it adds value for the reader.

In an ideal scenario, the workflow might look something like this: A marketing manager prompts ChatGPT to write a draft of an article on a given topic. Then the manager hands off the AI-generated rough draft to a skilled copywriter or editor who meticulously fact-checks, refines and humanizes the content. Once the polished piece is ready, a marketing assistant takes charge of publishing it.

When you don’t have processes in place for handling the potential pitfalls of AI-facilitated writing, it can undermine your reputation, destroy trust, and leave a bad taste in your customers’ mouths.

Militantly fact-check all GenAI output

I can’t stress this enough: You absolutely must check every line of AI output for accuracy.

The technology website CNET started using GenAI to write articles in conjunction with human editors. But 41 of the 77 AI-generated articles included errors the editors missed. CNET paused the program.

Ethan Mollick, a professor at the University of Pennsylvania's Wharton School of Business summed it up best by saying, “The best way to think about this is you are chatting with an omniscient, eager-to-please intern who sometimes lies to you."?

Check your output. Enough said.

Elevate your content by collaborating (not just generating) with AI

Smart companies aren’t treating AI like a content factory. They’re engaging and collaborating with it to create highly effective, accurate communication.

If you treat AI strictly as an output generator — and then don’t do anything to fact-check or humanize that content — you risk contributing to the current "tsunami of crap" online today (to borrow JA Konrath's phrase).

Like this article? You’ll love what we send out to our VIP email list. Get premium content from Horizon Peak and personal notes and updates from Jessica Mehring?when you subscribe to the Horizon Peak VIP List.