People are using ChatGPT for almost everything. Students are using the tool to complete their homework and assignments, and even professionals are using it in their jobs. However, ChatGPT has been spotted adding hidden watermarks in its text, which also get transferred to your documents when you copy-paste text directly. This is extremely alarming as it makes it easy to detect whether your text was generated by AI. In this guide, let’s learn how you can check if your ChatGPT-generated text contains hidden watermarks and how you can remove them.

Hidden Watermarks in ChatGPT-Generated Text

ChatGPT was found to be inserting hidden text watermarks to all responses using special Unicode characters that are not visible to the naked eye. However, when you insert the text in a code generator, you can spot these watermarks in the form of U+202F and U+A0 characters, which are Unicode formats of Narrow No-Break Space (NNBSP). Although it’s a generic text and not a special ChatGPT signature, this makes it very easy to detect that your text was generated by using ChatGPT.

When you copy and paste the response from ChatGPT to any other source, for example, a document, email, or even social media, these hidden characters are also copied as they are universal ASCII codes. Hence, it becomes necessary to analyse and remove such watermarks.

How to Check Hidden Watermarks in ChatGPT Text

SoSciSurvey is a free online tool that can detect and show you hidden watermarks in any text. Here’s how you can use it:

1. Open www.soscisurvey.de on your computer.

2. Paste the text generated by ChatGPT in the text box. Click on Show me the characters.

3. The website will highlight the hidden characters and watermarks in your text sample.

How to Remove Watermark from ChatGPT-Generated Text

Since the watermarks are Unicode characters, we need to use a code editor to remove them from our text sample. For this, we can use Sublime Text, a free code editor available on Windows, macOS, and Linux.

1. Download and install Sublime Text on your computer.

2. Paste the text copied from ChatGPT.

3. Go to Find and select Replace.

4. Now copy the special character watermark that you see in the text.

5. Paste it in the Find section, and enter a blank space in the Replace area.

6. Click on Replace.

Your hidden watermarks will be removed from the text sample. Now you can paste the answer in any document without worrying about getting detected for using ChatGPT.

Note: OpenAI says that the special characters in ChatGPT’s responses are not a watermark. These are just a quirk of large scale reinforcement learning, due to which the chatbot was accidentally using ACSII codes in blank spaces. The problem has now been resolved, and ChatGPT no longer inserts watermarks in its text and responses.

FAQs

Q. How to make ChatGPT undetectable?

You can make ChatGPT-generated text undetectable by using paraphrasing tools to reshape the sentences and add a natural flow to them. This will make it difficult to identify whether the answer was AI-generated or not.

Q. Does ChatGPT leave watermark in texts?

ChatGPT does not leave any special watermarks in its text. However, it may sometimes add unique ASCII and Unicode characters that make it easy to detect.

Wrapping Up

The watermarks in ChatGPT-generated text make it easy to spot the use of AI. By using free tools like SoSciSurvey and Sublime Text, we can detect these watermarks and remove them with simple steps, making it hard to detect. However, AI tools should be used with caution.

You may also like to read:

• This AI Tool Can Detect Scams in Photos, Videos and WhatsApp
• 3 Ways to Detect Fake AI Generated Videos Online
• 3 AI Tools to Find the Prompt Used to Create an AI Image
• Top 5 AI Object Removers Tested: Which One Works Best?

You can also follow us for instant tech news at Google News or for tips and tricks, smartphones & gadgets reviews, join the GadgetsToUse Telegram Group, or subscribe to the GadgetsToUse Youtube Channel for the latest review videos.