Proofreading is an essential step in producing high-quality written content. However, when relying on Language Learning Models (LLMs) like ChatGPT, the text can sometimes end up sounding too artificial or cringy after proofreading. LLMs, including ChatGPT, tend to use specific phrases and words that can make the text sound overly robotic or awkward. For example, ChatGPT might frequently use words like "elucidate," "delve into," "gauge," "pivotal," "burgeoning," "realm," "dynamics," "intricate," "nuanced," "interconnected," "interplay," "articulate," "delineate," "underscore," and "showcase."
In this guide, we will discuss some strategies to prevent this issue while still benefiting from AI proofreading tools.
Solution 1: Ask the AI to Avoid Cringy Words
The simplest way to tackle this problem is to explicitly ask the AI not to use cringy or overly formal language. Request the AI to replace those typical AI-sounding words and phrases with more natural alternatives.
This is my revised manuscript. I have made all the changes. I need your help to respond to the reviewer. Please write the response to conform to academic style: no spelling and grammar mistakes, clarity, conciseness, and overall readability. Don’t use contractions. Don’t use the following words and phrases: "elucidate," "delve into," "gauge," "pivotal," "burgeoning," "realm," "dynamics," "intricate," "nuanced," "interconnected," "interplay," "articulate," "delineate," "underscore," and "showcase."Solution 2: Provide Clear Guidelines
To maintain the original tone and voice of your text, give clear instructions to the AI. For example, ask it to focus solely on grammar and spelling corrections and to avoid making too many changes. Additionally, request the AI to indicate where changes have been made, making it easier for you to review and approve them.
Can you help me ensure that the grammar and spelling are correct? Do not attempt to polish the text; if no mistakes are found, please tell me that this paragraph is good. If you find grammar or spelling mistakes, please list the mistakes in a two-column markdown table. Put the original text in the first column, the corrected text in the second column, and highlight the keywords you fixed.
Example: Paragraph: How is you? Do you knows what is it? | Original sentence | Corrected sentence | | :--- | :--- | | How is you? | How are you? | | Do you knows what is it? | Do you know what it is? | Below is a paragraph from an academic paper. You need to report all grammar and spelling mistakes as in the example before. Paragraph: XXXSolution 3: Use Alternative Proofreading Tools
Another option is to use dedicated proofreading tools like DeepL, which allows you to choose different writing styles. You can ask the tool to make the text sound more natural while still correcting grammar and spelling errors. Also, some LLMs tool are better in terms of proofreading. For example, Mystral's Le Chat is better than Gemini in terms of proofreading.
Solution 4: Double-check the Proofread Version
Regardless of which AI proofreading tool you use, always double-check the proofread version of your text. This step ensures that the original meaning has not been distorted and that the tone remains consistent with your intentions.
By following these strategies, you can effectively use AI proofreading tools like ChatGPT without sacrificing the natural, human touch in your writing. Clear communication, appropriate tool selection, and careful review will help you maintain the quality and authenticity of your content.
Comments