As AI content creation tools become increasingly sophisticated, the ability to distinguish between human-written and machine-generated text has become a topic of intense discussion.
Enter the concept of AI identifiers – specific words, phrases, or patterns that are claimed to reveal the presence of AI authorship.
But are these identifiers reliable indicators, or simply a collection of assumptions?
In an attempt to seek clarity amidst the confusion, we gathered insights from leading digital marketing and content experts, asking them to examine the purported logic behind AI identifiers and to share their perspectives on the validity and limitations of these methods.
Their insights shed light on the realities of AI detection in the evolving landscape of content creation.
Read on!
Dawson Whitfield
One of the funniest identifiers of AI-generated content is the claim that it’s always “overly formal” or “robotic.”
Sure, maybe early AI sounded like a 19th-century butler writing an essay, but modern AI can drop slang, emojis, and even pop culture references like it’s trying to win the cool kid award.
Honestly, I’ve seen AI churn out content so casual and witty that I had to double-check that my team hadn’t secretly outsourced it to a stand-up comedian.
Conversely, the idea that “repetitive phrasing” is a foolproof giveaway feels a bit rich—have these critics read some human writing lately? We’re all guilty of overusing our go-to phrases, especially after coffee number three.
I think the real issue isn’t whether the content is AI-generated but whether it’s engaging, informative, and, dare I say, enjoyable.
Let’s stop being the content police and focus on what truly matters—connecting with the audience. AI or human, good writing speaks for itself.
Dawson Whitfield
CEO & Co-Founder, Looka
Dr Kevin Huffman
Among the suggested descriptors I find particularly funny is the concept of complicated vocabulary.
Here is why this is an ill-fitting marker for AI-generated content.
The use of advanced vocabulary can be something that appears in some AI writing, but not always. A large number of AI programs can mimic the habits of human speech even with simple words.
Conversely, human authors can deliberately use dense words in order to make themselves appear authoritative without really understanding what they mean. By looking at vocabulary level alone, we miss out on how humans communicate and what the evolving AI can do.
AI can sometimes falter at vocabulary, but this alone can be a deception.
The more holistic view – one that considers elements such as style, factual correctness, and general consistency — is better at flagging AI.
Dr Kevin Huffman
CEO & Founder, Ambari Nutrition
Lolly Spindler
I think my favorite identifier of AI-generated content is the word “tapestry.”
Considering how few people actually use that word when writing, it seems absurd that LLMs like ChatGPT spit it out so frequently!
I mean, honestly, if we’re not talking about an actual tapestry, are we really going on and on about “the company culture’s rich tapestry” or “transforming data into a tapestry of interconnected knowledge”?
Humans don’t talk—or write—like that!
Two or more things that are somewhat related don’t have to be a “tapestry”; sometimes, things can simply be related without being woven.
Jean Christophe Gabler
The theory that content free of errors was most likely produced by artificial intelligence truly bothers me.
Being a writer helps me to appreciate how difficult it is to guarantee that your work is error-free. Though I understand that even little errors can make your work more original.
When I wrote about health, the smooth writing was not because of an algorithm, but because I knew a lot about the topic and went over it many times.
I do not like it when this much information is missed because the writing is so good.
Jean Christophe Gabler
Publisher & Founder, YogaTimes
Claudia McKenzie
I do a lot of content and AI is a part of it now, as it has to be, but we are judicious about it and never just copy-paste. Some “identifiers” I spot right away are:
1. Title case for headers! I think that’s a dead giveaway of AI use. They were rarely used in the copy before and now AI has made them so much more common. Not to mention. most people would get them wrong, which AI does. So the perfect use and the abundant use are both AI giveaways for me.
2. Colons in titles. AI seems to love giving you titles with a colon in between, e.g.: AI vs Human Writing: Key Differences
3. Long dashes — ChatGPT loves to use the at least every couple of sentences when normal human writing won’t use it nearly as much
4. This one is more subtle but AI is rarely assertive. It will often predilect language that expresses opinions and possibilities instead of firmly stating anything.
Claudia McKenzie
Outreach Lead, Orange Line
Robin Samora
My pet peeves with AI include phrases like: in this digital landscape, not just … , ensure and leverage are overused… and there are so many ‘ing’s.’
Wordiness and flow are issues — you have to simplify AI copy to make it human.
The worst for me, is all the ing’s — it’s a dead giveaway!
These phrases and sentences drive me crazy because they’re so predictably AI, especially if you don’t know how to customize your prompts (repeatedly).
One fix is to seriously train AI on your voice — but you’ll still have to edit and of course fact check.
Robin Samora
Marketing & PR Consultant, RobinSamora
Rany Burstein
AI content today is not strictly for blogs, people use it for emails, business profiles and more.
We started to see more AI use on our site to create profiles. While they are of high quality, it also opens the fear that scammers from abroad who have poor English will use AI to create amazing and correct profiles.
Some easy giveaways include excessive emojis.
We’ve seen these before, but they are more common now. Perfectly aligned, bulleted bios with lots of Emoji.
There are also a lot of capitalized words for emphasis that don’t seem so natural.
I’ve also encountered this when looking for some freelance hires on Upwork. Amazing bios, communication via messages, but then in a real interview they barely speak coherent English.
Rany Burstein
CEO & Founder, Diggz
Stefan
Avoidance of Strong Opinions: I would point out that AI rarely takes strong stances on issues, often providing neutral or watered-down viewpoints.
For instance, instead of saying “This is the best solution,” AI-generated content may say something like “This could be a potential solution.” This overly diplomatic approach can feel oddly robotic and problematic in discussions that require conviction or passion.
Repetitive Sentence Structures: You see, AI content frequently exhibits repetitive sentence structures, such as always starting sentences with the same word or phrase.
According to certain critics, this can make the content sound monotonous and lacking in variety. This quirk can create a robotic tone that stands out but also leads to a peculiar critique: humans, too, can write repetitively when tired or rushed.
David Cooper
One term that I find particularly amusing is “unnaturally formal tone,” which is often utilized to describe AI content.
In my experience, language is quite personal and fluctuates greatly depending on emotional state, circumstance, and even individual personality. AI-generated information may be structured and formal, but it lacks the details of a person’s emotional undertones or the subtle ties that human language naturally possesses.
As someone who deals with language on a daily basis, whether in patient reports or treatment plans, I see AI’s precision as both a strength and a limitation, it lacks the deep, human touch that makes communication genuinely useful and important.
David Cooper
Strategic Advisor, Yung Sidekick
Debbie Moran
I find some of the common AI content identifiers a bit amusing, especially the ones like ‘too many adjectives’ or ‘unnatural phrasing.’
In my experience, these traits don’t always paint an accurate picture. AI can generate text that’s surprisingly human-like, making it tough to spot based on surface-level issues.
The real problem is that these identifiers focus on the obvious, when in reality, it’s often the subtle patterns in AI-generated writing, like how it handles nuance or structure, that give it away.
That’s what we should be looking for.
Debbie Moran
Marketing Manager, RecurPost
On behalf of the BoostMyDomain community of readers, we thank these leaders and experts for taking the time to share valuable insights that stem from years of experience and in-depth expertise in their respective niches.
BoostMyDomain invites you to share your insights and contribute to our authoritative publication. Reach a wider audience, build your credibility, and establish yourself as a thought leader in an industry that caters to every business with an online presence!
