In 2018, Keaton Patti went viral for making a bot watch over 1,000 hours of cheesy Hallmark holiday movies and asking it to write a script based on what it learned.
The results were hilarious — not only because the bot incorporated all the predictable holiday movie tropes, but also because of its writing style. It read exactly like what it was — a robot trying to sound like a human.
Fast forward to the present day. AI statistics show that it has officially become a part of the daily zeitgeist, advancing at world-record pace. Although AI chatbots are continuously improving and sounding more human, you don’t need to squint to see obvious red flags that the content was AI-generated.
Spoiler alert: we’re going to explain how to spot AI-generated text and how to tell if something was written by ChatGPT, Gemini, Jasper AI, Claude, DeepSeek, or other AI chatbots on the market without using AI detection software.
1. Opens With THE Cookie-Cutter Phrase
Google this phrase: “In today’s competitive” and see what happens. The results show a list of articles that eerily begin with the same or some slight variation of that phrase. Why is that?
Well, we gave ChatGPT a very generic prompt and here’s what it said:
Is that just how ChatGPT speaks? Glad you asked. We also gave Gemini the same prompt and:
When you see that cookie-cutter phrase in the first sentence, odds are, it’s an AI-written article.
2. Overuses Typical Words and Phrases
We combed through several articles and Reddit posts to find a curated list of words and phrases found in articles written by AI. The list was long and hearty, including words like “landscape,” “vital,” “crucial,” “additionally,” and especially, “delves.” In addition to users reporting the overuse of the word “delves,” and articles being written about it, ChatGPT’s social media acknowledges it.
We decided to go directly to the source and see if ChatGPT would give us a list. We asked ChatGPT to provide the top nouns, verbs, adjectives, transitional words, and phrases it overuses. The results were gold.
| Nouns | Verbs | Adjectives | Transitions | Phrases |
|---|---|---|---|---|
| Strategy | Leverage | Effective | Additionally | “In today’s fast-paced world” |
| Solution | Enhance | Strategic | Furthermore | “It’s important to note that” |
| Experience | Optimize | Comprehensive | Moreover | “A key takeaway” |
| Process | Implement | Robust | In addition | “As a result” |
| Collaboration | Facilitate | Dynamic | Therefore | “When it comes to” |
3. Repeats Writing Patterns
AI language models do what they know — meaning they rely on the datasets that they’re trained on. This results in using the same formula when building out paragraph structures. For example, in a three-sentence paragraph of AI-written text, the second sentence can sometimes feel like the only valuable one. The first and third sentences are fluff, usually repeating the general overview of the paragraph’s topic.
Additionally, many conclusion sentences have the same formula, mixing different Oxford comma lists together in the following structure:
“By [implementing paragraph topic], your [subject] [will/is a great way to] [general topic of the post].”
For example, “By following these customer service interview template best practices, your business can adequately prepare for customer service interviews.”
4. Provides Generic Explanations
When humans write, they often write from experience, providing specific details that help readers understand an idea or topic. AI language models have limited understanding or context for certain topics, which leads to broad, high-level explanations. This lack of nuance, plus easy access to generic information online leads to providing vague descriptions and filler words that add little value to the content.
5. Displays Strange Sentence and Paragraph Structures
AI likes to balance paragraphs, structuring them to be nearly identical in length. Humans focus on the informational flow and storytelling narrative rather than adding fluff to create paragraph uniformity.
Another red flag of AI-generated content is odd sentence structures. There are a few reasons for that:
- AI doesn’t always fully grasp the context of a conversation or topic.
- It’s using a dataset with awkward or complex phrasing.
- AI might translate certain concepts or phrases too literally.
In an effort to produce content based on the prompt, the AI model might copy language or regurgitate information it doesn’t completely understand — resulting in sentences that are grammatically correct but don’t quite make sense.
6. Contains Inaccurate Information
The internet can be a mesmerizing ocean of incredible and accurate information — and also some outdated, incomplete, and biased data. Incorrect information can even have serious repercussions, like giving harmful advice about eating disorders.
Based on your prompt (and how you word it), AI might access some of that inaccurate data to create an article with skewed or wrong information that makes readers raise their eyebrows. AI can also make mistakes called “hallucinations,” which means it makes up facts if it doesn’t have enough information on the topic. ChatGPT and Gemini often provide sources, but humans may not always confirm if the source is legit before using the content.
7. Shifts Voice, Tone, and Style
If you’ve ever been reading a blog post and get whiplash from the inconsistent shifts in voice, tone, and style, you might ask yourself: Was this written by AI? AI models can struggle to maintain a consistent tone, sometimes switching between overly formal to overly casual language.
For example, a blog post may be explaining different communication types in a professional tone when the next paragraph suddenly switches to overly conversational and says something like, “Hey, pal…ever had one of those days?”.
Human writers are naturally better at transitioning between ideas and keeping a consistent voice. AI-written content can be random or struggle to maintain consistent balance of voice, tone, and style, making it easily identifiable as AI-generated content.
8. Lacks Creativity
While AI can piece together existing ideas in ways that seem creative, it’s no match for the human imagination. AI’s creativity is limited to recreating what’s already out there rather than coming up with something truly original.
AI is better at providing straightforward answers without feelings, beliefs, or unique perspectives, like humans do. People tap into their own experiences, emotions, and cultural influences, resulting in innovative and relatable content.
9. Offers Outdated References
Seeing outdated references in a blog post might be more of a yellow flag than a red one. Depending on the context of the article, the reference might be appropriate (see: the holiday movie reference in the intro to this post). But if the article references, say, a product release date that was scrapped months ago, and the blog post date is current, it’s wise to proceed with caution.
Some AI models may not have real-time internet access so they have to rely on outdated knowledge bases. And if the knowledge base isn’t current, you’ll get outdated or obsolete results.
If you’re suspicious that an article was written by AI based on outdated references, you can verify the dates of any sources mentioned. Regardless if it’s AI-generated or not, the article may be unreliable until it’s updated.
10. Absence of Personality or Emotion
“I know now why you cry. But it is something I can never do.”
—T-800, Terminator 2
No matter how advanced AI becomes, it will never be able to capture the essence of human emotion and personality. Human writers naturally add these elements to their work, adapting their tone and style, and adding opinions, references, or personalized stories and anecdotes to the content, bringing it to life.
Articles written by bots often lack emotional depth and empathy, delivering the facts without connecting with the reader. Without the warmth or rich details that come from first-hand experiences, the content becomes obvious that it’s written by AI and also fails to keep the reader engaged.
Busted: Examples of AI Content Fails
It’s okay to use AI to enhance aspects of content creation, but it’s never okay to pass off AI-written articles as original content. Although Google doesn’t officially penalize AI-generated content, it may flag the content as unoriginal or low quality. Plus it just looks bad to readers and can damage your brand reputation.
Here are a few examples:
- Sports Illustrated was caught publishing AI-generated articles (and even used AI to generate the author and the authors’ bio photos).
- CNET’s reliability score gets downgraded to less than reliable after posting a flood of AI-generated articles.
- Gizmodo was called out quickly by Star Wars fans by posting an AI-written “Chronological list” article that wasn’t chronological and was packed with factual errors.
How To Use AI to Elevate and Accelerate Your Content Process
Using AI in your content marketing strategy can help human content creators enhance the writing process — but it’s all about finding the right balance. Here are a few ways you can use AI for good, instead of evil:
Enhance creativity and inspiration: Use AI to slay your writer’s block by helping you help spark fresh ideas through brainstorming prompts. This could be for blog writing, social media copy, email subject lines, or more.
- Create outlines: AI can help you create detailed outlines and map out your content to maximize the reader’s time-to-value.
- Find information gaps: AI can spot gaps in your content and suggest missing topics that would add more value to your article.
- Optimize SEO and content strategy: AI tools can analyze your content and offer tips on improving SEO, from tweaking keywords to refining meta descriptions.
- Adjust tone and readability: AI can suggest areas where your article could improve with readability and help shift your tone and style to suit your audience, from formal to conversational.
Pro tip: Use a tried-and-true AI prompt library to refine the AI chatbot’s results.
Human-first Content, Siege-Style
With AI becoming a mainstay in content creation, Siege Media continues to find new ways to innovate and supercharge our human content marketers. It’s all about using AI the right way.
AI-powered content marketing tools can provide our content marketing specialists with time-saving automations and extra inspiration to further enhance the quality of our content. Now that you know how to spot AI generated text, it’s time to start creating quality content of your own.
Let’s make that human connection and create something amazing together. Hit us up.
