The thing that sets the latest chatbots apart, and that of ChatGPT, is the use of natural language processing to generate human-like responses that are far more sophisticated than anything that’s come before.
Since OpenAI launched ChatGPT in November 2022, internet users have been experimenting with the conversational Artificial Intelligence chatbot in some entertaining and useful ways.
So far, we’ve seen examples of ChatGPT being used to compose music, write poems, create AI-generated content, give advice, and more.
It’s a sensationally clever tool with human-like language capacities and with seemingly limitless potential.
But ChatGPT also raises some ethical eyebrows. Specifically, over concerns about whether its content should be flagged for plagiarism.
TABLE OF CONTENT
Why People Are Concerned About AI Plagiarism
So everyone is on the same page here, let’s give a highly-simplified view of how ChatGPT works:
- The user types a prompt into the chatbox.
- ChatGPT uses its training data (roughly 45 terabytes worth) to form a response.
- You receive a response from the chatbot that is an amalgamation of what it compiles.
In most cases, there’s nothing here to get upset about. And for general purposes, AI is a great tool to assist with tasks and increase productivity.
The concerns arise when we apply ChatGPT’s immense capabilities to education and when it’s used to “cheat” work.
Universities have expressed concerns over their student’s potential misuse of ChatGPT and other AI tools.
While they are undeniably valuable for speeding up laborious research, the fact of the matter is that ChatGPT can write an essay or complete homework for you if you let it.
Another area of concern for people is the misuse of AI in paid work, for example, bloggers and journalists.
Rather than painstakingly researching and writing original copy, professionals could use AI to generate large swathes of text in a flash, saving time and effort.
A Common Sense Approach
Before everyone gets their knickers in a twist, it’s worth remembering that there is such a thing as AI text detection software.
These free-to-use tools will scan a body of text and indicate whether it’s AI-Generated writing or not.
Most of them these days are pretty accurate most of the time. Besides, for anyone who has actually used ChatGPT, there is a highly noticeable lack of personality in its responses.
Sure, it’s super clever and even mind-blowing at times, but a carefully trained eye can easily spot the robotic nature of its dialect.
Plus, there are a few limitations to ChatGPT, including:
- Character/word limits
- Frequent user capacity issues
- Errors generating responses to long prompts
- ChatGPT is only trained on data up to late-2021
- Can’t receive voice commands so everything must be typed
In addition to this, OpenAI, the ChatGPT creators,have already made their intentions clear on the issue of plagiarism, stating their “commitment to preventing plagiarism and other nefarious applications.”
OpenAI is alsointroducing a watermark that will be embedded into ChatGPT’s code, allowing systems to easily identify the content’s origins.
What Is ChatGPT’s Plagiarism Score?
Plagiarism scores are used to detect how much of a text is original and how much has been plagiarised using other sources.
Turnitin, an EdTech company on a mission to promote honesty, consistency, and fairness across education and academic research, is leading the charge.
They’re claiming they have technology that can detect AI-assisted and AI text-generating tools, such as ChatGPT.
Even with such plagiarism checkers, it’s virtually impossible to say what score the plagiarism score ChatGPT will get you.
Some reports claim ChatGPT has scored as low as 5% making it hard to spot these language models’ output.
A general guide is that 15-20% plagiarism is acceptable for essay work, as it naturally contains references to quotes and factual information but the end result will depend greatly on the content.
Is ChatGPT Plagiarism?
No, ChatGPT is not explicitly plagiarism. When you use AI tools like ChatGPT, you’re basically getting offcuts of information all stitched together in a somewhat coherent format.
Universities, teachers, employers, and others are concerned about the sheer lack of originality AI-generated content can promote and the implication for cheating.
Wrapping Up
Using AI writing tools or chatbots might not be outright plagiarism but in professional and educational spaces, it’s best to probably use them for their intended purposes as “tools”.
With new technology being developed every day that can detect AI-generated texts, this issue is unlikely to stick around for long but in the meantime, keep it original!