Penned

Signs of AI writing: 12 habits that give AI text away

Language models pick the choice that fits the most readers, so their writing repeats the same moves. These are the habits people notice first, each with a before and after.

The free Penned AI detector lists the habits it finds in your text next to the detector score, so you can see both at once.

1. Not X but Y

“It's not just a tool, it's a mindset.” The first half denies something nobody said, so the second half sounds bigger.

  • Before: Online learning is not just a backup plan; it is a whole new way of teaching.
  • After: Many universities now plan online courses from the start instead of treating them as a backup.

2. One-line closers

A short line after a paragraph that repeats it: “Let that sink in.” “That's the real win.” “The message was clear.”

  • Before: The new library opens at 6 a.m. and closes at midnight. That changes everything.
  • After: The new library opens at 6 a.m. and closes at midnight.

3. Staged run-ups

“Here's the thing.” “Let's dive in.” “Honestly?” The writer announces the point instead of making it.

  • Before: Is the course worth it? Honestly? It depends on your goals.
  • After: Whether the course is worth it depends on your goals.

4. Arguing with no one

“Don't get me wrong”, “I'm not saying”, “To be clear” before an objection nobody raised.

  • Before: Don't get me wrong, exams have their place. But coursework shows more.
  • After: Coursework shows more about a student than a single exam.

5. Sayings that sound deep

“At its core”, “The real question is”, “X is the language of Y”. They dress an ordinary point as a hidden truth.

  • Before: At its core, remote work is about trust.
  • After: Remote work only works if managers stop checking who is online.

6. Inflated significance

“Marking a pivotal moment”, “plays a crucial role”, “a testament to”, “the future looks bright”.

  • Before: The college opened a robotics lab in 2023, marking a pivotal moment in its history.
  • After: The college opened a robotics lab in 2023.

7. “Serves as” instead of “is”

“Serves as”, “stands as”, “boasts” where is or has would do.

  • Before: The hall serves as the main venue and boasts 400 seats.
  • After: The hall is the main venue and has 400 seats.

8. Lists of three

Ideas arrive in threes whether the meaning has three parts or not: “innovation, inspiration and insight”.

  • Before: The fest offers talks, workshops, and networking opportunities.
  • After: The fest has talks and workshops, and time to meet people between them.

9. Dashes everywhere

Em dashes joining every second clause. One dash is normal; a page full of them is a habit.

  • Before: The policy — announced without warning — affects every hostel.
  • After: The policy, announced without warning, affects every hostel.

10. Overused AI words

Delve, tapestry, showcase, seamless, vibrant, meticulous, underscore, realm. One is fine; several together are a sign.

  • Before: The essay delves into the vibrant tapestry of city life.
  • After: The essay looks at daily life in the city.

11. Bold labels on every point

Lists where each item starts with a bold label and a colon that the sentence then repeats.

  • Before: Performance: Performance improved with the new design.
  • After: The new design made the app faster.

12. Chatbot wrappers

“Certainly! Here is your essay:” at the top and “I hope this helps! Want me to make it longer?” at the bottom.

  • Before: Sure! Here's a revised version of your paragraph: ... I hope this helps!
  • After: Just the paragraph.

Keep the facts when you fix them

The easy mistake is cutting a fact along with the habit. When you remove a closer or merge a list, check that every name, number, date and quote is still there, and do not add details the original never had.

Readers and detectors look for different things

Fixing these habits makes text read better to a person. AI detectors also measure things you cannot see, like how predictable each word is, so a clean text can still be flagged. Check it in the detector after editing. The Penned humanizer rewrites for both: it scores every version with detectors, prefers versions without these habits, and checks the result against your original for facts. More on the detector side is in how AI detectors work.

Sources

This list draws on Wikipedia's Signs of AI writing, kept by WikiProject AI Cleanup, and on the open-source humanizer pattern list. The examples are our own.