Special offer. See more information about Outbyte and uninstall instructions. Please review EULA and Privacy policy.

Yes: studies have found certain words appearing more often in specific bodies of writing associated with large language models (LLMs). “Delve,” “intricate” and “underscore” are examples identified in research on scientific abstracts. But there is no universal blacklist of AI words, and a word choice cannot establish that an individual passage was written by AI.

Which words have researchers found appearing more often?

In their COLING 2025 paper, Tom S. Juzek and Zina B. Ward identified 21 focal words whose increased occurrence in scientific abstracts they considered likely related to LLM use. The paper’s abstract names “delve,” “intricate” and “underscore” as examples. The finding is about a measured change in a particular genre and corpus, not proof that those words are inherently artificial or that every LLM favors them equally. Read the paper abstract and record.

“Significant” illustrates why lists can become misleading. In an analysis of arXiv paper abstracts, Mingmeng Geng and Roberto Trotta found that the frequency of “delve” and several other publicized ChatGPT-associated words fell after early 2024, while “significant” continued to rise. This pattern indicates that usage changes over time; it does not make “significant” an AI tell. Read the study abstract and record.

Why might these words become common?

Researchers have not established a single cause. Juzek and Ward report no evidence in their analysis that model architecture, algorithm choices or training data caused the pattern. Their model tests were consistent with reinforcement learning from human feedback (RLHF) contributing, but they describe the causal question as unresolved and note limited transparency around model development. That is suggestive evidence, not a settled explanation.

What’s actually slowing this PC down?

Pick the symptom - the matching free tool is one click away.

Special offer. See more information about Outbyte and uninstall instructions. Please review EULA and Privacy policy.

Word frequencies in published writing can also reflect what people do with model output. Authors may prompt, select, revise or rewrite generated text. Geng and Trotta interpret the decline in some publicized words after early 2024 as consistent with authors adapting their use of LLMs once the pattern drew attention. Their study therefore describes changing human-LLM writing practices, not a fixed fingerprint of machine-generated prose.

What other writing research adds—and what it cannot establish

A 2024 Scientific Reports study comparing AI-generated, AI-revised and human-authored application materials reports in its indexed abstract that AI-generated documents used a smaller vocabulary and repeated favored words. That is a different genre and set of writing conditions from scientific abstracts, so it should not be merged into a single ranking of “most AI words.” Read the study.

Rank #2
SKATAN Letter Writing and Envelope Addressing Guide, fits a Wide Range of Letter Paper and Envelopes
  • Multi-function in one: letter writing guide and envelope address writing guide.
  • Multiple slots: contains 1/6", 1/4", 1/3", 4/9" and 5/9" slots for different letter sizes.
  • Large size :8 inches wide for all types of letter paper and envelopes.
  • A wide range of applications: wedding invitations, scrapbooking, making, CARDS, etc.
  • Professional quality: good toughness, odor-free, soft and durable, translucent plastics.

A 2025 PubMed-indexed study examined more than 15 million biomedical abstracts from 2010 to 2024. Its authors’ excess-word method suggested that at least 13.5% of 2024 abstracts had been processed with LLMs. This is an estimate produced by that method for that corpus—not a direct count of disclosed AI use or a prevalence figure for all writing. See the PubMed record.

These findings differ in genre, time period and measure: some examine changing word frequencies, one reports vocabulary diversity and repetition, and another uses excess vocabulary to estimate LLM processing. They do not establish a universal set of AI words or a common baseline for every kind of writing.

Special offer. See more information about Outbyte and uninstall instructions. Please review EULA and Privacy policy.
Rank #3
Resholder 10 Writing Guides 20 Low Vision Sheets 2 Thick Pens for Seniors
  • Set Includes: writing guide kit letter writing guide set includes a writing guide set, including 10 writing guides, 20 low vision papers and 2 thick-tipped low-vision pens for elderly people with low vision
  • Various Styles: 10 template guides are divided into 5 styles, 2 sheets each, which can help low-vision users master five practical writing formats: letters, personal checks, envelopes, business envelopes and signatures
  • Rich quantity: 40 low-vision papers, thick black lines and bright white paper, suitable for low-vision and visually impaired people, size about 8.5 x 11 inches/21.6 x 28cm
  • Low-vision aids: bold pen tips can produce clear and easy-to-read line widths, which are very suitable for low-vision and visually impaired people
  • Practical and considerate: this rich set of low-vision aids can meet the practical needs of low-vision users and can be given as gifts to those in need to express care
Independent reader supportYour contribution helps us test, update, and keep practical guides available for everyone.Support on Ko-Fi

Can a word tell you that a passage was written by AI?

No. Corpus-level overrepresentation means that a word appeared more often across a defined collection than expected or than in an earlier comparison; it does not identify the author of a specific sentence. People have always used words such as “delve” and “significant,” and an author may edit model output—or use a model without retaining its characteristic phrasing.

Geng and Trotta’s findings make detection harder, not easier: if writers change or edit output after a phrase becomes conspicuous, a word-based signal can fade. A suspiciously repetitive or formulaic style may prompt closer review, but a word list is not a reliable AI detector and should not be used as proof of authorship.

Best Value
Sale
Magic Practice Copybook, Reusable Writing Practice Book, for Preschool Kids Age 3-8 ​Calligraphy 9.44in×6.29in(5 Books with Pens)
  • Magic Practice Copybooks are much larger than other copybooks on the market. The package includes 5 magic practice copybooks for Kids (Addition,Multiplication,Numbers , Alphabet, Drawing) + 1 Pens + 10 refills
  • The Magic Practice Copybooks has three-dimensional groove design, The magic calligraphy paper is thick, strong, and durable, not easy to smudge and not easy to stain; the pen holder is made of silicone, which is 100% safe for children; the print is clear and clean; it is really comfortable and great for Calligraphic Letter Writing.
  • The matching refills use magic ink,handwriting will automatically fade and disappear within 30 minutes after writing, great for repeated practice. It is a Must Have handwriting copybook set for your kids learning calligraphy.
  • The kids can use their magic pen to follow the paths and draw different shapes, great for improving their writing skills. They can also master English handwriting skills and write beautiful English words through training with the reusable writing board tracing books.
  • Magic reusable calligraphy copybook set is the perfect gift for your kids ages 3-8 years old. It is suitable for all kinds of holiday gifts, such as Christmas, New Year, birthday.If you have any problems, please contact us at any time, we will reply within 24 hours.

How to read claims about “AI words”

  • Check the corpus. A result about scientific abstracts does not automatically apply to emails, essays or application materials.
  • Check the timeframe. A word’s frequency can rise or fall as models and authors’ habits change.
  • Check what was measured. Frequency, repeated favored words, vocabulary diversity and estimated AI processing are different outcomes.
  • Separate association from cause. A word becoming more common in LLM-associated writing does not by itself explain why it changed.
  • Do not treat ordinary vocabulary as a tell. Context, baseline usage and the writing genre matter more than a single word.

Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.