As a scientific editor, I spend most of my time reading scientific manuscripts. Recently, I started noticing some manuscripts use words and phrases that felt oddly formal or just...off.

You know, like how the word ‘delve’ has suddenly started to appear everywhere?!

Some research and conversations with industry colleagues confirmed my suspicions: the culprit is the now widespread use of AI large language models, such as ChatGPT.

(When I refer to AI in the rest of this post, I’m specifically talking about these large language models and text generators, rather than applications of AI in scientific research.)

AI is a big change that academia hasn’t quite adjusted to yet. There are both pros and cons to the use of AI in scientific writing or any type of written or creative work. And, yes, there are ethical issues.

However, I also know that most scientists are already using AI to help them proofread or rephrase their text. When used correctly, some AI apps can handle these tasks well.

For those of you who are using AI to check or polish your text, I thought it would be useful to put together a list of ‘AI Red Flags’ to help you avoid the suspect words and phrases that might tell your reader, supervisor or journal reviewer that you’ve been hitting the prompts in ChatGPT.

If you’re not using AI, then this list will help you avoid using any Red Flag phrases or words unintentionally.

How did we identify the list of AI Red Flags?

Our Customer Support Specialist, Marjorie, scoured the web to find the words and phrases that people identify as tell-tale signs of AI-generated text. We reviewed numerous articles, blog posts and forums to create a comprehensive list of Red Flags in any type of writing and text, not just in scientific papers.

We also looked at several scientific reports, including Monitoring AI-Modified Content at Scale: A Case Study on the Impact of ChatGPT on AI Conference Peer Reviews and ChatGPT “contamination”: estimating the prevalence of LLMs in the scholarly literature. These studies tracked the trends in the use of specific words over time, showing some massive increases in the use of specific words since ChatGPT entered the picture.

Overall, we identified 305 words and 43 Red Flag phrases that are considered to be signs of AI in any type of writing. Here’s some of the words and phrases we found:

You can access the full list of AI Red Flags by entering your email in the form at the bottom of this page, and we’ll automatically grant you access to the Google sheet.