Is ChatGPT plagiarism Free or does it plagiarize?

It’s no secret that ChatGPT has become an important content writing tool for many people. However, there are some questions of plagiarism concerns regarding ChatGPT’s output.

The first thing to know is that ChatGPT’s GPT language model did not invent or create any original new content. Rather, ChatGPT has various ways of explaining what is already known or can be found online.

This is how the concept of generative AI works. Generative AI provides responses to prompts in its own unique way. So technically, there is always some degree of borrowing knowledge from other sources to provide an output.

Some questions that come up when writers and content creators start using ChatGPT for content creation include: Is ChatGPT plagiarism free? Do we own the content? Can we monetize the content? In this article, we will answer all of these questions.

Is ChatGPT plagiarism free

ChatGPT has been trained by consuming an immense amount of text from the internet, books, and more. This means everything that ChatGPT knows and how ChatGPT works is based on knowledge and information it acquired from somewhere.

Plagiarism means presenting a piece of work, ideas, or sources as your own without consent from the original author, and this is the case with ChatGPT. Nobody knew that their work, website, or blog article would be used to train ChatGPT. 

You’re probably thinking I could ask ChatGPT to provide sources, right? You can ask, but we’ve seen examples of ChatGPT making up fake sources because it was just asked to provide a source. Unless your version of ChatGPT has access to the internet, it won’t be able to tell you where the information is coming from. 

I just asked ChatGPT to explain the theory of relativity and then came up with this prompt:

When prompted for a source, ChatGPT starts listing books about the theory of relativity. Although those books do exist, ChatGPT insists thatit can’t give you the source of the knowledge and that these are just recommendations. 

ChatGPT plagiarism and OpenAI’s terms of use

To answer this question, let’s have a look at OpenAI’s terms of use:

“Input and Output are collectively “Content.” As between the parties and to the extent permitted by applicable law, you own all Input. Subject to your compliance with these Terms, OpenAI hereby assigns to you all its rights, title, and interest in and to Output. This means you can use Content for any purpose, including commercial purposes such as sale or publication, if you comply with these Terms”

This means that while we own the content that ChatGPT generates, we must also ensure it’s plagiarism-free and respects local laws and regulations.

Let’s take a look at a couple of random scenarios to understand:

Scenario 1

You’re a student working on an essay about the impact of car batteries on the environment. You provide ChatGPT with a prompt asking for an informative paragraph on that topic.

The AI generates a response that seems to make sense to you. You copy it directly into your essay without any modifications or attributions. You do not do any fact-checking or try to find where the information comes from.

In this case, even though you own the output generated by ChatGPT, directly incorporating it into your work without proper citation would be considered plagiarism.

Scenario 2

Suppose you’re a content creator tasked with writing an article on the benefits of meditation. You use ChatGPT to help generate some content and receive a well-structured paragraph on meditation’s mental and physical advantages.

While the output isn’t copied verbatim, you slightly rephrase the AI-generated text and integrate it into your article without acknowledging the source. Even though you’ve rephrased the content, not giving credit to ChatGPT or the original authors (if you know them) for the initial idea and structure still constitutes plagiarism.

NB: Those scenarios were generated with the help of ChatGPT and adapted afterwards.

If you really want to use ChatGPT for whatever task you have to do, and the plagiarism will be checked, here’s what you should do:

• Use AI as a starting point and not a tool to give you the final result.

• Cite your sources (whether it is ChatGPT or another source)

• Fact-check the output and list the sources you used to fact-check

• Paraphrase, rewrite, and add examples in the final text.

• Add your story, your “why”, your reasons for talking about the topic you’re covering.

• If using ChatGPT is part of the assignment for research purposes, say so and document everything as you would document research for another topic.

Do AI detectors work?

This is a big question that scares students all over the world. Can an online tool determine if something was written by AI accurately?

To find the answer, I infiltrated two Facebook groups of teachers using ChatGPT to help them teach, and the answers I found during my research were not the ones I expected.

Here are some testimonials that I got from members who wished to remain anonymous:

“I’ve tested this with several AI detectors. I was able to run AI written and personally written items and get a false reading on each sample every time”

“I am yet to try an AI detector, from what I’ve read, it’s more along the lines of a very faulty, advertised marketing shenanigan”

“AI detectors are unreliable. They are just a marketing tool”

So, how about we test that out?

Let’s take the first part of this article and run it into six AI detector tools. It should come up as negative, as this is original content. Afterwards, I will test some output from ChatGPT (the output about the theory of relativity from earlier).

I asked ChatGPT to rephrase the content to pass AI detection tools. I didn’t remove any formatting from ChatGPT. I will highlight the times when the detector tool was wrong (red), right (green), and somewhat wrong (orange).. Let’s see the results:

It seems clear that some of those tools aren’t able to detect AI content properly. Results from AI detection tools do not constitute ultimate proof of anything, and it is up to each institution to decide whether online tools can be used to accuse someone of plagiarism.

As many teachers suggested in the Facebook groups, it’s better to have a discussion with the student than accuse them of plagiarism based on the result provided by an online free AI detection tool.

Round Up

When we look at ChatGPT plagiarism concerns, it’s pretty clear that it’s not intentionally giving plagiarized material. Its responses can unintentionally resemble existing content due to its training data.

This will become a topic of discussion at some point in the future, and we might see OpenAI including a sort of watermark to indicate to the reader that what they have their eyes on comes from ChatGPT and may contain unintentional ChatGPT plagiarism.

As always, the proper use of AI rests on us, and we must ensure there is no misinformation, plagiarism, or copyright infringement in the content we use ChatGPT to create.

Thanks for reading, and I hope you have gained a better understanding of ChatGPT plagiarism questions. If you want to take your knowledge and skill with ChatGPT to the next level, why not check out our other ChatGPT content on Automation Switch?