Anytime robots do human things, we get anxious. And we bring that anxiety to Google. When people search “AI copy,” they also research:
So... will AI (artificial intelligence) copywriting tools actually steal your content marketing job?
“The future doesn't have to suck,” he told MarketerHire. “AI is not here to take your job or to come after your family.”
Founded in 2020, CopyAI is a newer player in the AI copywriting space. Its founder Paul Yacoubian brought on Emal to ramp up its marketing in February as the first hire. Considered one of the best AI copywriting tools for marketers, the AI content generator is making waves on Twitter and lists on Google.
Emal’s on a mission to demystify AI copywriting tools and let people know they’re just that — tools. Here to help, not take over.
To emphasize this, CopyAI didn’t give their tool a cutesy human name, like some of their competitors. Instead, it left AI in the name. According to Emal, the CopyAI team wanted to make it clear their product is not a r̶e̶p̶l̶i̶c̶a̶n̶t̶ substitute for a human writer.
It’s artificial intelligence. But we can get more specific than that.
How do AI copywriting tools work?
Like all current AI copywriters, CopyAI relies on Generative Pre-trained Transformer 3 (GPT-3), the third-generation language prediction model created by Elon Musk’s AI research lab OpenAI.
It has 175 billion parameters — 173.5 billion more parameters than the previous GPT model.
AI copywriters like CopyAI are GPT-3 dressed up — basically, an intuitive skin over OpenAI’s technology that makes it more usable for non-tech people. Other AI writing tools powered by GPT-3 include:
- Conversion AI
- Shortly AI
If GPT-3 is Steve Wozniak, CopyAI is Steve Jobs. You need both.
“It formats specific prompts and gives you great outputs,” explained Emal. “If you're a social media manager and you need to write a great tweet, or you need to write landing page headliners... there are specific tools for each of those use cases.”
AI content generators create text from human inputs — think keywords, outlines and short briefs — using an open-source AI. They can create multiple rough drafts of social posts, subject lines, product descriptions and more at warp speed — but even according to their biggest fans, they’re a supplement to human creativity, not a replacement for it.
AI copywriting tools are faster at creating rough drafts than humans. Even 150 WPM typists can’t pump out seven blog intros in less than 30 seconds.
Not only can AI spit out more words, it can dream up alternative angles, helping “expand your thought process,” according to Emal.
He clarified that it’s not a copy-paste solution, though. “After you use the tool… it's expected that you should modify.”
“After you use the tool… it's expected that you should modify.”
Much like great blogs start with great outlines and clever creative starts with clever briefs, the quality of the output is directly correlated to inputs. Here’s the product disclaimer from CopyAI:
Even GPT-3’s co-creator recognizes it’s far from human. “[T]he hype is way too much,” tweeted OpenAI co-founder Sam Altman.
Still, just for fun, I asked CopyAI to write an intro for this blog. I started by inputting a brief:
Title: “End of the (human) copywriter?”
Description: What AI content generators mean for writers
In a few seconds, it gave me seven intro options, ranging from one to three sentences.
Here’s the first (and best) option:
The rise of the machines isn’t just a science fiction fantasy anymore. Software bots are writing scientific papers, helping you choose which Netflix series to watch and now, in the latest horror story for copywriters, drafting company emails. This week research from startup Aisin AI revealed companies are using content generators to fuel their outbound communications. The software helps scale these messages to hundreds of thousands of customers worldwide — and lets businesses cut costs.
Honestly, not bad.
One potential problem: It’s not obvious, but the Aisin AI research referenced here is meant to be a placeholder.
When I searched for “Aisin AI” in Google, it was nowhere to be found. Naturally, I couldn’t find the research referenced, either.
You can see how this could cause problems if I actually used that sentence in this article. The lead is so original, it’s fictional in places.
Is AI content always original, though?
I ran the paragraph above through Quetext’s plagiarism checker, and it turned up no issues — which Emal says is typical.
“The content you're creating is pretty much 100% new,” Emal said.
“The content you're creating is pretty much 100% new.”
CopyAI generates text based solely on user inputs, he explained. It doesn’t scrape the web for content — or even snippets of content. In other words, it doesn’t pull from Google like ClearScope or Fraise.io, two popular SEO optimization tools.
That’s why CopyAI isn’t great at web research — but also why you can trust the marketing copy to be fresh.
The 5 biggest benefits of AI content generators (according to human copywriters)
AI content generators are far from perfect. They can't research or churn out polished copy — and they all have ties to Elon Musk. But they still make writing content easier in a few ways.
1. They reduce your weekly word count.
“We don't consider how much we actually have to [write] in our day-to-day life,” said Emal.
Between writing and editing, a single person can create millions of words in a month. Like, more than five million.
2. They help with brainstorming.
And kill writer’s block.
3. They speed up the writing process.
“I see scrappy startups using [AI content generators] to move fast,” predicted Brooklin Nash, head of content at Sales Hacker and small business owner of SaaS content agency Nash Content Consulting. They can help small businesses and marketing teams hit ambitious deadlines.
“I see scrappy startups using [AI content generators] to move fast.”
4. They shoulder repetitive writing tasks.
In the same way Zapier automates tedious workflows, AI writing tools automate the tedious parts of content creation, according to Emal. Things like:
- Writing product descriptions for similar SKUs
- Creating slight variations on the same ad
- Generating proposal and email templates
5. They turn non-writers into passable writers.
In addition to assisting professional writers, AI content generators, Emil says, can “help people that aren't good at writing copy” — or don’t do it for a living.
The 3 best use cases for AI-generated content
You probably shouldn’t have an AI content generator writing letters from your CEO. The longer, less structured and more heartfelt content needs to be, the less AI can handle it. But when it comes to quick-hit text, it’s great at generating lots of options — and finding infinite new ways to repeat itself.
Product descriptions are the most popular use case for Copy AI, Emal said. CopyAI typically provides seven short descriptions. If you like one, you can ask it for more.
Here’s how Elvis might have used an AI content generator to write variations on product descriptions for his favorite shoes:
Content creators also use the product “a lot for blog intros or headlines,” according to Emal. Most content writers don’t use the first headline they write, so instantly generating 20 more slight variations is appealing.
Here are some headline ideas CopyAI gave me for this blog (using the current title):
Social media posts
Social media posts are another popular use case, according to Emal. Especially, when you need to come up with hundreds of clever hooks a week.
From social post hooks to video titles to YouTube description intros, AI content generators equip social media managers with social copy tools they didn’t know they needed.
For instance, here’s an Instagram caption CopyAI suggested I use to promote this story:
You can tell that CopyAI wrote it, and the ending is a little abrupt — but hey, it’s a starting point.
The roles, careers and industries most obsessed with AI copywriting tools so far
“Key people that use our tool are social media managers, e-commerce store shop owners,” said Emal. “And agencies love us, because we make it so easy to do things on a massive scale, for their customers.”
Here are their most common use cases:
E-commerce owners: Product descriptions.
“If you have 100 products on your store that is so tedious to write out a story and a product description for all of those,” explained Emal.
“If you have 100 products on your store that is so tedious to write out a story and a product description for all of those."
Instead, AI content generators can produce those unique product descriptions for you — in as little as 30 seconds. But, before you go and just paste them on your site, be sure to do an edit on every individual one.
Copywriters: Proposals, emails, and intro or headline ideas.
As Kaleigh suggests, many copywriters are taking to AI content generators as a way to generate new ideas and squash the blank page forever. Sometimes, all you need as a copywriter is a few ideas on the page and something to get your started. You’ll go back and edit it all together anyway, likely changing that AI writing beyond recognition (i.e. to sound like more natural language, or like that copywriter’s own voice).
Students: Resumes and cover letters.
Everyone hates writing about themselves. Tweaking and updating the same cover letter and resume is particularly dull. AI copywriting tools offer templates, so you’re not staring at a blank screen or getting too deep into your life-story.
Agencies: Ad copy.
“If you're writing 1,000 ads a day, it can be kind of helpful to have some help on that for sure,” said Emal
“If you're writing 1,000 ads a day, it can be kind of helpful to have some help on that for sure."
Emal is right. Paid digital advertisers write thousands of ad copy variations a day — testing different creative with different messaging. As long as algorithms on Instagram, LinkedIn and the like continue to dictate what creative and copywriting is shown to whom, and when, advertisers will continue to test every possible ad copy variation. Machine learning can make Facebook ad, Google ads and the landing pages they drive to way less painful, and even help generate results far faster.
Social media managers: Hook ideas and social media captions.
Social media is instant, so speed and quantity are critical. At larger companies, social media departments require multiple captions, titles, hooks, memes, videos, and more –– every day. Copywriting tools can help social teams keep up with the demand and create more bite-sized content, faster.
The drawbacks of artificial intelligence for content creation
CopyAI clients use AI to take on quite a few copy tasks — but can it do them well? Depends on who you ask.
“It is a very, very good copywriter,” said Emal.
“It is a very, very good copywriter.”
But GPT-3 leaves something to be desired, according to copywriters and content marketers.
AI content tools still struggle with creativity and context. Again, GPT-3 doesn’t pull from the internet, so it’s extremely limited when it comes to pop culture references and connections.
To test out CopyAI, Brooklin Nash, head of content at Sales Hacker, asked the tool for a few headlines about… itself.
It gave him these blog title options:
“Are any of these good? Maybe it'll help spark an idea, but it seems like it's creating more work instead of saving time,” concluded Nash.
“[I]t seems like it's creating more work instead of saving time.”
Here’s what felt off to Nash about the AI-generated headlines:
- They don’t sound quite human. “A blog about how you can create” just isn’t what a real person would call this.
- They’re too straightforward. “A blog that talks about how copywriting can help your marketing” is… off-puttingly meta.
- They’re wordy. The first headline uses “Copy AI” and “about” twice – that’s too wordy for precious headline real estate.
“This isn't a knock against CopyAI, just seems like GPT-3, in general, is a lot more limited than folks are making it out to be,” said Nash.
The question everyone is asking: Will AI replace copywriters?
Not the good ones. That was the consensus between Emal, copywriters and content marketers.
“Heck no,” Nash told MarketerHire. “There's a creative bent that's not anywhere close to being replaced with a $99/mo tool.”
“There's a creative bent that's not anywhere close to being replaced with a $99/mo tool.”
We’re pretty sure an AI content generator couldn’t write Laura Belgray’s homepage copy for her Talking Shrimp freelance business:
Belgray’s homepage is a highly human creation, between the “me want” button and the footnote citing “an unscientific but totes accurate study.”
Like Nash, Kaleena Stroud, copywriter at ISDIN, isn’t worried about AI copy tools. In college, she read Roald Dahl’s story, “The Great Automatic Grammatizator,” about (spoiler) the pitfalls of algorithmically generated writing.
It convinced her that “nothing can ever replace voice,” she said.
Eddie Shleyner, founder of VeryGoodCopy and former lead copywriter at G2, is more split, though.
“Eventually — unfortunately — I think [AI] will affect some writers,” Shleyner told MarketerHire — technical writers and journalists in particular.
“Eventually — unfortunately — I think [AI] will affect some writers.”
Only because GPT-3 is very good at writing about facts in a “natural, familiar way."
AI has already made its way into the Washington Post newsroom, where it acts as a hybrid content management system, helping The Post cover hundreds of races on Election Day.
Shleyner thinks copywriters are safe from AI infringement, though.
“Copywriters… are usually writing about emotions, which is nuanced and tricky. Just Google ‘GPT-3 pickup lines’ if you don't believe me,” he said “AI has a long way to go before it can appropriately channel the human condition.”
(A sample GPT-3 pickup line: “Are you a candle? Because you’re so hot of the looks with you.”)
Like Shleyner, conversion copywriter Jackie Quiring isn’t stressing about job security. AI copywriting tools can compete with bottom-tier writers, but not the cream of the crop.
Sales copy that “converts like crazy” will always be rare, says Quiring.
“I know I'm biased since I write words for a living,” he said. “But… I don't believe in [GPT-3] for copywriting, marketing and other creative processes.”
“I don't believe in [GPT-3] for copywriting, marketing and other creative processes.”
Conclusion: AI takes over, but not in the way you think
In theory, instead of asking a junior copywriter to write 20 taglines in an hour, creative directors could start using AI copywriting tools to generate 200 taglines in minutes.
The problem? AI doesn’t understand intention.
“Personally, I'd rather have 20 thoughtful tags in front of me than 200 random thoughts,” said Shlyener.
“Personally, I'd rather have 20 thoughtful tags in front of me than 200 random thoughts.”
Emal also clarified that copy tools can’t form relationships. So, humans can always cling to that.
“AI isn't going to be able to talk to the specific customers of my company, get what their feedback is, be able to do a full audit, and actually write the copy.”
More likely: Tools like CopyAI will automate away writer’s block.
“AI will change the way copywriters ideate, the way we brainstorm.”
“AI will change the way copywriters ideate, the way we brainstorm,” predicted Shleyner. “Getting started might become less painful.”
But identifying the best ideas and connecting them in an “efficient, elegant way… will still fall on the copywriter,” said Shleyner.