The Research Which Stopped Me Worrying About AI Taking Over My Writing Job

I’ll be honest, when my colleague sent me an email saying that the coschedule title analyser had given my title a low score and I should improve it, I was not impressed. 

How can a computer know more than a writer about what is and isn't going to persuade someone to click on a title? I was proud of the title, I liked it.

Her email went onto tell me that I should analyse all of my titles with this machine and improve them for better readability.

I dismissed the advice.  I think I spent about 30 minutes checking out the tool and decided what can it know and moved on.

Six months later, AI technology is popping up more and more in content creation. It’s starting to get my attention and I’ll be honest, does make me slightly concerned.

If a machine can write my content, what am I here for?

I still rely on content writing as a form of income.

I imagine a lot of other writers and copywriters are having similar questions and fears.   How long will it be before a robot is doing our job.

Research a little bit online and your fears will not be squashed! By the looks of it in the media we are all going to be out of a job next week.

But, how much of that is realistic, or just hype?

It’s in the best interest of the AI technology companies & their investors to create a lot of buzz and “positive sentiment” as an AI robot might say. They want companies to feel like if they don’t jump on this technology straight away they will be left behind in the dust - their competitors speeding ahead as they adopt AI into their strategies.

So, I did some research of my own.

Going into the research adamant that a robot was going to take away my job, and worse my only true passion. A writer writes because that’s what they are on this planet to do. Take away the reason to write, what has a writer got left?

However, after completing this research I feel better - and I hope after reading this you might too.

This isn’t going to be a scientific study I’m afraid. I’m a writer, not a scientist.

I’m also obviously bias because I don’t want robots to take away my job, or my passion.

But, I don’t think those things mean I shouldn’t have a voice on this matter. If anything we need more voices on the other end,

For this unscientific study of AI technology & copywriting I focused on a tool, Atomic AI, which is pitched as a tool to help marketers write better. It does this by telling you if you are writing for the right level of audience and scores you accordingly.

I wanted to look at its scoring system, relative to how many shares a piece of content got.

The premise being, if a piece of content scores higher, then it should be shared more. That’s what  the tool is helping writers with.  To write for their audience better, make a deeper connection in order for them to share or engage with the post, which is what every marketer wants as a result from their content.

For this study, I’m going to look at the same website.  Because, as we know the size of an audience has a direct impact on how many shares an article will get.

Yes, so does a million other factors.  This is not scientific, remember.

I’m also going to use the blog content of a website I admire and respect for their copywriting.

Kopywritingkourse.com

As with every single blog out there, some content does better than others.

Some content gets a higher level of engagement from the audience.

This is what we are always trying to improve on.  To identify what it is that makes our content great + try to replicate that greatness for future posts.

So, let’s see if content that got a lower level of engagement on Kopwritingkourse.com, gets a lower score on Atomicreach.com

The Pages I Chose From The Homepage Of The Blog >

1. https://kopywritingkourse.com/airbnb-listing-description/ - 73 shares

2. https://kopywritingkourse.com/create-advertising-that-sells-david-ogilvy/ - 223 shares
​
3. https://kopywritingkourse.com/billboard-advertising/ - 116 shares



Now Let’s see what each one gets as score from Atomic.

1. https://kopywritingkourse.com/airbnb-listing-description/ - 73 shares

2. https://kopywritingkourse.com/create-advertising-that-sells-david-ogilvy/ - 223 shares

3. https://kopywritingkourse.com/billboard-advertising/ - 116 shares

Let me explain what these screenshots from Atomic mean.  

So, for each piece of content you have to select a readability level.  After playing around with it all afternoon I’ve realised that more of the online content, even on the larger blogs try to write for a reliability level “knowledgable”.  

Hot Tip: This is actually in itself an important takeaway.  Initially I assumed that all of the top blogs in my industry were focusing on writing for an academic audience. It turns out, knowledgeable is the level they are all going for.   In a sec I’ll explain what this means!

There are 5 levels in Atomic.

1. General
2. Knowledgeable
3. Specialist
4. Academic
5. Genius

The basis of this tool is that it will tell you if you are writing at a level your audience can understand.  So, I select knowledgable for these three pieces of content.

Then, the tool will score the piece based on the setting.

The way it calculates it is basically how advanced are the words you are using.

For example, 

Original: Located in north Austin, this bulletin targets traffic heading north towards Pflugerville, Round Rock & Georgetown, avoiding traffic on I-35.

"avoiding”, is set as knowledge level specialist, one level above where we are aiming.  So I can click the word and choose one more suitable for my audience.  Changing these words will increase the scoring of the piece overall.

When I click "avoiding"

I am told it can be switched to shunning, shirking or squashing.

Improved: Located in north Austin, this bulletin targets traffic heading north towards Pflugerville, Round Rock & Georgetown, shunning traffic on I-35.

Clearly this word doesn’t improve the sentence, it just doesn’t feel right. It feels way too dramatic.  I SHUN THE TRAFFIC - who shuns traffic??

Anyway, that gives you an idea about what the readability level does and how it applies to the text.   I don’t think the tool is particularly great at coming up with accurate synonyms.  HOWEVER, it is useful to know at what level you are writing at.

We are always trying to write in an approachable way.  To cut the jargon and get straight to the point.

But how do you actually know that this is how you are writing?

I believe this is one way to find out, so just for this reason I suggest you log on for free and check it out.  Further than that though, I fail to see how this tool is useful long term.

Good for competitor research, and for checking your writing style, but is this robot going to take my job.  NO. At least right now I'm safe.

The second thing this tool does is gauge emotion.

Emotion looks at the language used and gives a score from calm to exciting.   The premise being that more emotional language gets shared more often.

To be honest, I don’t know how it works this score out. Previously, tools would categorise the types of words used and then assign either positive or negative to those words and create the score from there.

But that caused issues, because sometimes a seemingly negative word can actually be used in a positive way.

For example the sentence:

The skateboarder executed a toe grab 360 and it was sick.

The words ‘executed’ and ‘sick’ are usually used in a negative way, but in this sentence the statement is positive.  The tools would not be able to accurately know this.

With Atomic AI this sentence is scored as ‘exciting’.

Exciting can be either negative or positive.

That's all I really have to say on it at the moment.  I don't really understand how this is particularly usful aside from the fact that I'll know if my writing is sending my audience to sleep or not.

Let’s move onto the scores that Atomic gave these three posts and see if the score is reflective in the number of shares it received.

Post 1  > Shares 73 > Atomic Score 74
Post 2 > Shares 223 > Atomic Score 74
Post 3 > Shares 116 > Atomic Score 76

The scoring is pretty similar across all three posts. Post 2 was the post about David Ogilvy, which if you read it, is actually a word for word copy of one of their copywriting adverts.  Shouldn’t this be the benchmark for copy, being that it was written by the most famous advertising agency of all time, and a living legend?

In fact, Post 3 got the highest score.

What I’m unclear about on this tool though is why. What happened to those two points, does that really matter? Or is this tool more useful for writers who have a lot of learning to do?

Maybe I put in text that was too high quality.

In summary, after doing this research I am not worried about my writing job (just yet anyway). There is so much hype about these tools.  This is just one, there are more on the market. 

I don’t want to completely discount all of them. Some might be useful.  But what I do want to highlight is the need to go on a try them for yourself before getting so concerned and worried.

So far I’ve been disappointed by what these tools can do, which makes me feel a whole lot better.  Hopefully you are too.
​
I'd love to know your thoughts in the comments.  Have you tried any of the new AI tools out? What are your thoughts?