Writing is an art. From the classic novels that have stood the test of time, to the copy you see in advertisements for brands like Coca-Cola, words have power and meaning. One reason they have this power is because there’s a human behind them. A human being with unique thoughts, experiences, and a story to tell that may just relate to your own. 

Large Language Models (LLMs) have made the idea of mass-producing written content appealing. Trained on massive datasets, these algorithms have ensnared the minds of businesspeople eager to cut costs within their various industries by cutting out that pesky human element and relying solely on machine-made content.

That’s not the best way to write effective content and can end up being downright harmful to your business, but that doesn’t mean there isn’t a place for AI in your content creation process.

Map the Road Ahead
While we don’t recommend using AI to create your content from scratch, it can be helpful when deciding what content to make. 

If you’re planning your monthly content calendar and you want all your posts to revolve around a certain theme, a tool like ChatGPT can break that theme down into subcategories that you can use when creating social posts, blogs, or landing page text. A simple prompt saying something like, “On the subject of back-to-school preparation, generate four blog post ideas that cover the different aspects of the subject” can give your writing direction. Even if it doesn’t break it down the way you’d like, you can still take inspiration from the output.
One… Two… Three… Break it Down!
You have a lot of information you want to convey, but an email isn’t the place to do it. Enter an LLM. 

Too much text in an email or social post is a surefire way to lose your reader’s interest. Those venues are meant to be hooks that bring people into the larger works and aren’t meant to convey all you have to say. You can feel all the text you want to condense into ChatGPT and say, “Summarize the following in one paragraph 3-4 sentences long, ” then paste all the copy on your landing page. It’ll create a paragraph that covers all the main points and leaves out all the nitty gritty. If they want to know more, they can head to your page!

Amazon is already doing this with customer reviews for products. They have an AI taking all the reviews and “averaging them out” to one review that covers the main points of what most consumers are saying. 
“Don’t You Take That Tone With Me!”
Here’s where things start to get a little tricky because we’ve already talked about how you shouldn’t let these LLMs write for you, but sometimes you have exactly what you want to say but just don’t quite know how to say it. Even copywriters have their off days! Feeding your writing into something like ChatGPT and then saying “Rephrase this more optimistically” or, “Make this sound more professional” can make all the difference when you are trying to convey ideas in a certain way or to a certain audience. You can create two different versions of an email and see which one plays better for your target audience. However, we do still recommend putting the results into your own words because…
Did a Machine Write This?
By and large, people can still tell when an LLM wrote your words for you. Some just know what to look for in terms of tone and structure, and in more extreme cases, people will be taking your copy and feeding it into AI detection software to determine if an LLM was used. 

Using these tools is not without controversy. As we said before, LLMs are trained on massive sets of data and that’s how it’s able to string together words in a way that sounds vaguely if not convincingly human. The problem is that the data it’s being trained on is content created by human beings who did not permit their work to be used in the creation of these algorithms. This puts an understandably dark mark on LLMs, as many consumers are aware of the impact this is having on actual human beings. Copywriters (and artists, but we’ll get into that in a future article) are already finding it harder to secure jobs in their field as businesses opt for the cheap way out when it comes to producing content. Relying solely on AI to write your words for you will seem like a slight against real human workers, and can create distrust with your audience. This is why we suggest writing everything yourself and not just relying on ChatGPT to do it for you. 

Knowing what you want to say before involving AI is also important because…
It Just Makes Stuff Up
There was a time when we at SLX experimented with AI for writing our copy, and we did this for the same reasons that other organizations have tried the same: increased efficiency. We decided to try it out on something easy and have it write a step-by-step process on how to do something in Pardot (MCAE).

Reader, it did not go well.

While it did write everything very confidently (why wouldn’t it?), it got every single step in the process wrong. It didn’t know the correct names of the tools, it included information that was long out of date, and when it came to numbers and stats, it just pulled them out of the air. It was all meaningless. If someone followed the process that the AI came up with, they wouldn’t have just done it wrong, they wouldn’t have been able to do it at all.

This is because–despite its name–Artificial Intelligence isn’t actually intelligent. It’s not able to learn the way a human mind does, and it’s just stringing things together in a way that the algorithm says sounds human. It can get 90% of your content right on a good day, but that 10% is all it takes for your audience to lose faith in your messaging. You cannot and should not rely on it as a resource for actual information. Maybe someday, but absolutely not today.

Needless to say, we basically had to rewrite that step-by-step article from the ground up, effectively taking more time to correct it than it would have taken a human being to write it. A dramatic decrease in efficiency.
One tool can’t do everything. If you’re a construction worker heading to a job site, you’re not just packing a hammer in your tool bag and calling it a day. Everything has a purpose, and LLMs are a tool just like any other. They can’t do it all.

They do have their place, though. They can bring nuance and some good tweaks to your copy and marketing that can benefit your org in the long run, you just need to be smart about it. Don’t get lazy and don’t start replacing your human writers! Learn to use the technology available, and trust your writers with your messaging. At the end of the day, the best way to connect with human beings is by using your humanity.
Share Article:

AI Is Nothing Without People Behind It

We are those people, and we’re here to help!


Current Waitlist For All New Subscribers: 21 Days