And the practical tips you need to make sure your copy shits all over the AI-generated stuff
Since AI tech bro’d its way into our lives in 2023, copywriters have kept a wary, watchful eye on what it’s got up to. A wise move, because in what feels like a nanosecond 28% of big-name companies have started using AI for marketing and content development.
This figure seems high given how recently AI rolled out to the mainstream, so I get why some of my colleagues are nervous. But not all of us are freaking out. First up, we’re confident human creativity can hold its own, because,
“AI is great for some things but it’s not great for everything. The difference between AI-generated copy and truly creative copy written by humans is clear. We just need people paying for copy to value it enough to choose real copywriters over robot ones.” – Vikki Ross on LinkedIn
And as this comment on one of my blogs attests, “…the world needs talented copywriters like you to cut through the AI crap.” (Cue blushing and awkward thank yous.) It seems at least some of my readers agree that AI is problematic.
Second up, for those of us who have been working in copy for a decade or longer, we’ve already seen this kind of thing.
Before AI copywriting, there were content mills
The challenges AI presents to copywriters and businesses aren’t anything new. Let me take you back to 2011 and the start of my copywriting career.
E-commerce businesses and online publications commissioned me to write their copy and blog content. They trusted that my three years as a journalist on a UK lifestyle mag equipped me for the job. My most frequent brief from these businesses was, Help us rank number one on Google.
In the early noughties, the SEO landscape was different. The tactics to get to number one back then would make today’s SEO strategists scream and run in circles. Crapping out content purely for chasing rankings and metrics was the norm. Thinking about and talking to the person you were writing for, not so much.
I felt like I was morphing into a content monkey, tasked to create stuff that would rank, rather than genuinely inform or entertain.
This approach was completely at odds with my journalistic values and background. It made zero sense that empty, often puerile, SEO tick-boxing content would outrank decent articles. Although nothing good could come of this, it turned out to be true. Because content mills proliferated.
Content mills. The stack ‘em high, sell ‘em cheap approach to copywriting
Content mills gave rank-hungry businesses what they most desired. A large volume of content, stuffed with keywords and crammed with links. The result? Thanks to the search algorithms operating at the time, a seat at the top ranking table.
Content mills lured writers to their platforms with the promise of regular work, for which they were paid an absolute fucking pittance. To offset their meagre returns, writers were promised, ‘The bigger your profile, the bigger your paycheck.’ In reality, achieving either was tough. (So I’m told.)
What actually happened was that writers and businesses entered a price race to the bottom. To make bank, writers were forced to focus on quantity rather than quality. Fuck the imperative to create anything wildly different, original, entertaining, or even readable. As long as X keyword appeared Y number of times and Z product was mentioned, it was job done.
The similarities between content mills and AI copywriting generators
For me, AI has no place in writing actual copy. It suffers from the same quality and readability issues as some of the guff generated in content mills. So the AI tosh that has to be ‘cut through’, as my blog commenter put it, is nothing new. It’s also easy to see the connection between AI and content mills when we think about how AI has learnt to ‘write’.
Large Language Models (LLMs) and AIs learn by ‘ingesting’ a huge amount of information. A lot of this content will have been scraped from copy typed by writers who were racing against the clock. These writers (for whatever reason) have prioritised quantity over quality.
In short, AI has inevitably learnt from production-line articles that were churned out before search refined its ranking criteria to encourage … better.
Honestly, we’re not battling a new issue here. 🤷
What copywriters are seeing and doing with AI copy
While some writers (me included) flat out refuse to use AI, others are figuring out how to work with it. Accepting that businesses are turning to AI and finding ways to integrate it into their workflow or build it into their services.
Some copywriters do this by offering to humanise your AI content or fact-check its claims. Compared to commissioning a writer to create a piece of copy from scratch, asking them to ‘de-bot’ some AI copy can be a cheaper option.
I get the attraction. And while I haven’t been asked to do this, my colleague and editor, Mary Cameron, has.
For balance and perspective, I asked her which businesses are using AI to write their copy and what she has to do so it appears human.
“When it comes to who is sending it, it’s a real mix. Some of my small business clients that I regularly write or edit blogs and newsletters for are trying out AI. They either write an entire draft or create briefing notes for me,” Mary says. And she believes they use the tool because they’re “curious about AI’s time-saving potential.”
The other people sending her AI copy are marketers needing to “get to grips with AI on behalf of their own clients.” They may use AI if they’ve been briefed to try something new—a style of writing that’s not usually in their wheelhouse. They ask AI to do the bulk of the legwork, add their own touches, then send it to Mary to edit.
Ultimately, Mary says “They want a zombie exorcism. Can I ensure no one will know AI has had its robot mitts all over their copy? Can I ensure that it speaks to and enthrals their audience and builds their brand?”
Making tweaks and edits to AI copywriting is important for two reasons.
- Search engines are already cracking the shits about the overuse of AI generated copy.
- People are getting good at spotting AI-written copy. And there’s backlash—it comes off as disingenuous and lazy.
But there’s something deeper that nags me about AI copy. And it’s this.
Why would I bother to read something that you couldn’t be bothered to write?
I don’t know who said this first, but for me, it gets to the crux of the issue.
I get why business owners are using AI. It can save time. (Assuming you feed it sufficiently succinct and accurate prompts.) But is there a way to reclaim time and achieve content you’re so proud of that doesn’t have to go through a machine, a writer, and an editor? How do you make your copy shit all over the AI stuff?
4 writing techniques to bury the crappy bot copy
AI copywriting is rife. And most places using it are lazy AF, which is good news for the rest of us. It means it doesn’t take much to create copy that’s shits all over Arseificial Intelligence. Here are four (relatively) simple things you can do to set your copy apart.
1. Be original. Gather your own stats, facts and quotes
Chatbots and AI generated-copy scrapes information that’s already available. If you gather original stats, facts and quotes then you have something that bot-generated copy doesn’t.
You might get this information by
- interviewing a subject expert
- running a short survey on your social media platform
- sending questions to your email list.
Generating your own original sources isn’t always possible. When this is the case, commit more time to the research part of your writing process. Dig out little-known sources that aren’t being used everywhere else. Even looking beyond the first 20 search results can yield these.
2. Sound like a human. Use familiar language
Unlike the LLMs you and your great, big human brain get out and about the world. You have the huge advantage of hearing people talk.
I’m a fan of this great anecdote about legendary advertiser David Ogilvy. He’d (kinda) punish copywriters who used complex language by ordering them to take a ride on a Greyhound Bus. If you’ve ever travelled by Greyhound you’ll know that these are long bus trips. You’re stuck in a metal box for hours with the great and good of humanity. Listening to the other passengers chatting is unavoidable.
If Ogilvy’s writer heard the words they used in their copy? Fair enough. If they didn’t, which was usually the case, he’d order a rewrite.
So, until bots start booking seats on Greyhound (and even after they do), you too, can eavesdrop on conversations around you. Listen for local dialect, idioms, accents and so on and thread some of those phrases into your copy to separate it from the tin-earred machine stuff.
3. Be surprising. Write boldly, differently and do the ‘unexpected’
AI writes using patterns. So it commonly:
- includes ‘unusual’ or jarring word choices
- omits emotions and personality and has a flat tone
- repeats itself and sticks to formulaic sentence structures
- makes sweeping generalisations without much evidence.
Best the bot with some creative rule breaking so your writing is lively, punchy and packs a surprise or three. Maybe, make up your own similes and metaphors (ones your audience will ‘get’ ), and chuck in a super short sentence or a long lyrical one.
I’m seeing a push for more humour in copywriting. AI sucks at being funny and can’t do sarcasm. You can. Or, if you kinda can but kinda can’t then get onto copywriter Lianna Patch. She’s been driving this idea for years with her Punchline Copy business and more recently the Conversion Comedy course, while Dave Harland is launching his humour-filled copy course, Write The Funny.
4. Be you. Build trust using your voice and stories
People buy from people. It’s a trust thing. Which is why defining your business tone of voice and consistently writing in that style is going to help you.
Now, I can already hear the AI crowd saying, Yeah but Rose, you can train AI to write in your brand tone of voice. True to a point. Although the subtlety and nuance of your personality will always be beyond it. Besides, only you can give yourself the upper trust hand by enhancing your copy with original images or videos.
Another thing AI can’t do? It can’t share personal anecdotes or opinions. Because it’s not lived. It’s never felt the sun on its face, experienced the satisfaction of hitting the snooze button, or known what it feels like to hold its first born child. You have. Sharing personal experiences or ideas relevant to what you’re writing about is going to set you apart.
Link up with like minded #teamhuman folk
Separating your copy from Arseificial Intelligence will stand you in good stead with search engines, your audience, and your conscience. I fully believe that when it comes to writing and creativity, anything the machines can do, we can do better. And I’m not the only one feeling like this.
The hashtag #teamhuman is growing amongst creatives who either aren’t using AI or don’t use it as part of their creative process. And courses, posts and emails are showing how humans just do better. You might want to take a look at:
- Vikki Ross’ post on X showcasing lines she doesn’t think AI could have ever written
- Eddie Shleyner using a Good Will Hunting monologue as a relevant lesson in overcoming AI
- Felicity Wild’s brazenly named, Writers Against the Algorithm—a community of copywriters who discuss how to navigate and thrive in a changing industry.
- the Copy CPR workshop run by Ami Williamson who is a self-professed AI-obsessed word fiend—this would be good for small biz owners.
So where is content headed in the age of AI?
In 2022 Google introduced the Helpful Content update, aiming to reward people-first content. This update recognises content and copy created with the reader in mind that hasn’t been manipulated to satisfy search engine rankings.
Basically, the world’s biggest search engine wants people and businesses to increase the quality of the content they’re publishing. It’s another step away from the content churn we’ve seen previously and, in turn, a disincentive for people ripping stuff straight from AI generators and posting it.
On one level, I’m all for this. It’s the argument I’ve been putting forward for years—that the interests of the person you’re writing for are at the heart of all good copy. Combine this update with the rise of some of the techniques I’ve discussed, and it feels like there’s a shift to encouraging business blogs to align more closely with journalistic practice.
And yet, I also sense there’s something darker lurking in the background.
As you know, everything from AI is scraped from existing sources. Eventually (if not already) content is going to become very boring and samey. To counter this, new content is needed. New content that we’re being encouraged to create but that will, eventually, end up being read and analysed by LLMs and AI chatbots.
Google owns and is developing one of the planet’s most powerful AI minds. Google is rewarding us for generating new content. Go figure.
The conspiracy theorist in me is wondering, is this just Google’s way of getting even more information, new information, different information to feed into their LLM to take it to the next stage?
It’s a big idea. And a bit of a scary one. So while we wait to see how Google’s plans for world domination do or don’t play out, let’s keep creating original, entertaining, reader-focused content.
The inspiration for this article
This article was inspired by points I put together for a podcast, namely an episode of Small Business Sweet Spot, where I chatted to host Barb Davids about how to differentiate your content from AI.
Listen to the whole episode here or search Small Business Sweet Spot to listen on Spotify.