An epidemic of content marketing is currently underway. Websites, blogs, and social media feeds are overflowing with AI-generated content, which makes consumers increasingly dubious of the material that companies post. 

Customers are quick to identify what they perceive to be indications of AI-generated copy, casting doubt on what is genuine and denouncing companies who use ChatGPT and similar products excessively. 

Copywriters and content marketers are questioning their job as a result. All of a sudden, we’re questioning if a well-known phrase or a writing style we’ve been using for years could make our work seem fake instead of genuine. 

In this post, we’ll discuss whether or not people can recognize text produced by artificial intelligence (AI), how to write more humanely, and which so-called "AI words” you might wish to block in ChatGPT. 

Key takeaways 

Can people really tell if content is AI-written or not? 

Many individuals have strong opinions about being able to identify AI content if you look at LinkedIn and other social media sites. Nevertheless, they lack confidence in their own capacity to do so. 

Only 47% of the 63 instructors in engineering, economics, and social sciences who were asked to assess brief passages of writings produced by humans and artificial intelligence were able to accurately identify AI-authored content. The detection rate fell to barely 20% when the AI-generated material was written professionally. 

In other words, even highly educated experts struggle to differentiate between literature created by AI and that created by humans. 

However, worries about AI content are still growing. 

This distrust is evident in forums and social media sites. If you spend five minutes on Reddit, you will see hundreds of people asking whether something is real or artificial intelligence-generated. Over 4 million people frequent the topic-focused subreddit every week, appropriately named r/isthisAI. 

These discussions become more difficult as generative AI improves. A growing number of readers and, consequently, prospective clients are searching for clues that enable them to assess if information is reliable, genuine, and authored by a real person. 

Is it really possible for people to distinguish between content created by AI and that created by humans? The jury is still out, in my opinion. 

Additionally, we do something pretty intriguing: we ask AI to tell us what is AI and what isn’t when people can’t trust their own ability to do so. 

Why AI content feels like AI 

The explanation for why AI content seems like AI is clear. The reason is that it adheres to formulaic patterns. 

Trillions of words written by actual people in books, websites, and articles are used to train AI algorithms. In order to inform its writing, AI gains knowledge of common terms, copywriting strategies, frameworks, and structures during this training. 

Because AI content uses specific words, phrases, or punctuation, people believe it is identifiable. However, similar characteristics of AI writing are equally valid methods employed by actual content creators. 

The reason why AI content feels like AI is because: 

AI struggles to add intensity to its writing, but it is extremely good at sounding correct. Because it lacks real emotions and life experience, it attempts to imitate the writing style of seasoned authors, which makes it sound prescribed and predictable. 

How to spot the signs of AI-written content 

Before we get started, it’s important to remember that a lot of indicators of AI-generated text are actually indicators of generic corporate writing. It’s not surprising that AI writing frequently sounds familiar since AI learned from previously published internet content. 

Look for patterns that indicate something lacks depth, such as random claims, clichéd phrasing, or prefabricated emphasis, rather than viewing individual words as “proof” that it was authored by AI. 

Steer clear of generalizations that lack substances: 

AI-generated information frequently relies on broad assertions that sound authoritative. Make your writing stronger by supporting important statements with facts, figures, examples, or logical arguments that lend credibility and provide readers with something worthwhile to take away. Try saying “content marketing generates 67% more leads than traditional marketing channels” rather than “content marketing is essential for modern businesses.” 

Substitute specifics for general adjectives: 

AI-generated text frequently relies on terms like cutting-edge, forceful, and inventive rather than offering significant detail. Whenever feasible, substitute specific facts that show value for descriptive fluff. Try “our platform reduces customer onboarding time from five days to just two hours” in place of “our innovative, cutting-edge platform delivers powerful results.” 

Eliminates superfluous clichés: 

AI frequently uses overused business terms that don’t really add anything to the content. While occasionally employing them is OK, go over your writing and eliminate any stock phrases that don’t enhance clarity or add meaning. 

Break predictable patterns: 

AI writing frequently depends on formulaic contrast, overuses the rule of three, or uses repeated sentence structures. To produce a more organic and captivating rhythm, change the length and structure of your sentence. There is nothing wrong with these frameworks in and of themselves, and if you want me to stop employing the rule of three, you will need to take it away from my conversion copywriter. However, your writing sounds too robotic if each paragraph has the same cadence. 

Go beyond superficial insights 

AI frequently provides general explanations of subjects without contributing complexity or uniqueness. Incorporate concrete examples, relevant advice, real-world experiences, or data that provides readers with genuinely helpful insights to bolster your text. 

Steer clear of overly dramatic language: 

AI frequently uses exaggerated terms like "revolutionary," "game-changing," and "unlock" to make concepts seem more amazing than they actually are. By using clear, measured language that appropriately conveys the value you’re presenting, you may foster trust. I learned from my copywriting coach not to use the verb “unlock” unless I’m discussing something that can actually be unlocked with a key. When revising information produced by AI, that is a helpful guideline to remember. Exaggeration should be avoided since it might damage credibility and confidence. 

Check for bias and confirm facts: 

AI is capable of producing false statements, erroneous figures, nonexistent sources, and biased or one-sided viewpoints that mirror trends in its training data. Instead of acknowledging that it is ignorant, it will frequently come up with a response and deliver it with total assurance. It makes it impossible to compromise on fact-checking content created by AI. If not, you might publish inaccurate information. 

What Google wants from content in the age of AI 

Given how much effort marketers spend producing content for search engines, it’s critical to comprehend Google’s perspective on AI-generated content. 

Fortunately, Google has started time and time again to say that it does not automatically punish the use of AI in marketing. They are more interested in the user experience your material provides than in whether or not you used AI. 

Google’s objective is still the same as it has been for many years: rewarding information that is beneficial, helpful, and fulfilling to the searcher. 

The problem at hand is that generative AI makes it very simple to scale content creation without contributing anything novel. 

Google recommends posing a straightforward query to yourself in its April 2026 guide, "Optimizing your website for generative AI features on Google Search”: 

The best content according to Google: 

To put it another way, don’t simply copy what is already available online. Add knowledge, experience, and creative ideas. 

You’re generally headed in the correct direction if someone finds your article, reads it through to the end, and feels like they’ve learned something. However, you’re offering visitors very little reason to trust, remember, or return to your website if your content simply restates what dozens of other websites have already said, AI or not. 

Additionally, Google advises being open about how you employ AI. If automation or artificial intelligence plays a significant part in your content strategy, be transparent about its application and adhere to Google’s recommendations for using generative AI on your website. 

This isn’t really about AI versus stuff authored by humans. It has to do with producing valuable content. 

How to make your content sound/feel human 

Readers are becoming more wary of anything that seems formulaic or generic due to AI-generated material. 

Making your material more human by adding traits that AI finds difficult to imitate—experience, knowledge, personality, emotion, and perspective—is the greatest approach to stand out. 

Add experience and expertise. 

Talking about your own experiences is one of the simplest ways to make your writing more relatable. 

Your content is exclusive to you because of your knowledge and experiences. Google’s EEAT -T framework, in which the two E’s stand for Experience and Expertise, is likewise satisfied by doing this. 

Provide readers with proof that the author is knowledgeable about the subject matter. In light of this, it’s crucial to incorporate quotes into expert-led material. 

Write with personality. 

Personality is one of the main things that AI-generated material loses. According to that reasoning, what distinguishes your material is your personality. 

Your content’s minute details reveal your personality: 

Either get to the point or write with emotion. 

The fact that AI frequently falls somewhere in the middle—neither succinct nor captivating—is one of its greatest flaws. 

Instead, decide on a definite course of action: either get right to the point or narrate an engaging tale with real emotion and experience. Simple refrain from packing pages with meaningless words. 

Write how you speak. 

Reading your writing aloud is one of the easiest methods to make it more relatable. Rewrite it if it doesn’t seem natural in conversation. 

Compared to AI’s frequently predictable patterns, writing the way you talk produces a more genuine rhythm. 

Don’t try too hard to sound human. 

Writing by humans is inherently untidy. However, it all gets a little too obvious if you try to force that messiness. Additionally, I’ve witnessed individuals swing excessively in the opposite direction. 

I’ve observed authors: 

What is an AI detector and how does it work? 

Many people employ AI detection tools since it might be challenging to determine whether information was created by AI. These technologies determine the probability that text is AI-generated by analyzing elements such as sentence patterns, writing style, and structure. 

Reseacrh presents a more complex picture, despite some claiming accuracy rates as high as 99%. Research has demonstrated that AI detectors are disproportionately biased against non-native English writers and often create false positives. Stanford researchers discovered that 61.3% of articles written by non-native speakers were mistakenly marked as AI-generated. 

I’m not arguing that AI content detection techniques are worthless. They certainly have a place. However, you shouldn’t regard them as conclusive evidence. A tool’s conclusion that ChatGPT wrote the Declaration of Independence should generally be regarded with caution. 

Step-by-step guide: Follow these instructions to create an AI buzzword blacklist. 

Words like “seamless,” harness,” and “delve” were in use long before artificial intelligence (AI). There is nothing wrong with the words themselves. It’s the regularity, consistency, and setting in which they are employed. 

Nevertheless, there are some terms and phrases that are so closely linked to AI-generated material that you should check them before publishing. 

I’m not advocating for a permanent ban on them. However, if you incorporate AI into your content creation process, implementing a personalized blacklist can help cut down on recurring AI clichés and promote more intelligent writing. 

Consider this more as a list of terms that merit more examination than as a list of words that are prohibited. 

Common AI buzzwords 

Common AI phrases 

Common AI conjunctions 

Common AI conclusions 

How to get ChatGPT to blacklist / avoid these words 

By providing custom instructions, as seen in the ChatGPT example below, you may lessen the possibility that your favorite AI tool will use specific phrases. Every few months, think about updating this list by (even) allowing AI to look it over and recommend new terms! 

Steer clear of corporate jargon, overused AI expressions, and clichés like "It's not X, it's Y.” Words like “delve,” “leverage,” “harness,” “unlock,” “seamless,” “robust,” “transformative,” and “revolutionary” Should only be used when absolutely essential. Give detail, actual examples, first-hand knowledge, and plain English precedence over marketing jargon. Steer clear of generalizations, needless paradoxes, and ambiguous conjunctions. 

Alternative option: Improve AI-generated content by creating a custom GPT 

Creating a Custom GPT with more stringent authoring guidelines is my preferred method. 

Before giving your GPT a list of the prohibited words, inform it that it is “strictly forbidden from using any of the following words or phrases in your responses.” It is more likely to pay attention to your blacklist as a result. 

Your blacklist of terms and phrases can be uploaded as a document or entered into the instructions box located under your GPT’s Configure tab. 

You can provide more instructions on how you want (or don’t want) AI to write if you’d like. To reduce pointless repetition and constrasting assertions, you could, for instance, give the following instructions: 

Repetitive patterns like the rule of thumb should be avoided. Make sure to change up the length and structure of your sentences. You should refrain from rephrasing concepts or making needless opposing comments. Avoid using contrasting comments like "It's not X, it's Y” or other needless clarifications. Write clearly and concisely. 

Next, use your freshly created Custom GPT whenever you publish text using ChatGPT. 

How to get Claude to blacklist these words 

By doing the following, you can get Claude to put specific words or phrases on a blacklist: 

Steer clear of corporate jargon, overused AI expressions, and clichés like “It’s not X, it’s Y.” Words like “delve,” “leverage,” “harness,” “unlock,” “seamless,” “robust,” “transformative,” and “revolutionary” should only be used when absolutely essential. Give detail, actual examples, first-hand knowledge, and plain English precedence over marketing jargon. Steer clear of generalizations, needless paradoxes, and ambiguous conjunctions. 

Final Thoughts 

No words or phrases are intrinsically “AI” in nature. The combination of general assertions, recurring patterns, and a lack of nuance, opinion, or emotion is what gives content an AI-generated vibe. 

As AI advances, readers will be more concerned with whether content is actually helpful than with how it was produced. The information that sticks out will be unique, experience-driven, and actually beneficial—qualities that Google also promotes by emphasizing originality and competence. 

While AI may mimic writing styles, it cannot reproduce your personality, viewpoint, or life experience. Your biggest advantages over the competition are still those.