Submitting AI-generated work without proper acknowledgement to the source is considered plagiarism at most institutions. Nevertheless, such content could be considered technically original in that AI doesn’t necessarily copy directly from another student’s work or source.  However, it does not reflect a student’s own analysis, understanding or critical thinking of the subject. Some universities classify AI-generated submissions as contract cheating, which is the same category as hiring someone to write an essay for you.  Students have argued that using generative AI technology to assist with their assignment may not seem like plagiarism because the answers come from an ethereal, complex computerised database and not any real person’s work. However, it’s not that simple. Even though AI-generated answers don’t appear to violate someone’s specific intellectual property, using them in an assignment without proper citation is academically dishonest because it is not your work.  This is especially true if it’s hard to tell the work apart from your own words which may or may not also be sourced. In academic settings it’s expected that students will appropriately source anything they’ve used in their research to develop their arguments, and this should therefore include any content generated using AI tools.

Originating from the Latin word ‘Plagiarius’, meaning “kidnapping,” plagiarism started as a metaphor during the Roman Empire for “kidnapping” one’s words from another. We now recognise it as a form of literary theft, where one person’s ideas are directly copied and passed off as a student’s own. It is considered a violation of integrity and ethics and is frowned upon in nearly every field and specifically in academia.  Plagiarism can lead to serious consequences. For example, a medical student may have no understanding of a specific procedure because they have never researched around it or read about it themselves, or a psychology student writing about serial offenders without having any understanding of what that constitutes.  Both have far-reaching consequences for any student once employed in the ‘real’ world.  Using AI probably means they know nothing about the subject other than the basics. Additionally, the act of plagiarism leads to serious issues of credibility, damaging the reputation of the university and questioning any findings the student may profess to have found. In higher education students who are caught plagiarizing can fail their assignments or exams, as well as facing suspension or dismissal from the course. There can also be financial consequences further down the line which can stem from accusations of copyright infringement.

Since AI draws on and paraphrases from many different available online sources it is difficult to determine how much of it is plagiarized from a specific source.  AI-generated content draws from existing, real-world sources and the algorithms that are the foundation of AI tools analyse existing content from across the internet, from academic and scientific journals, blog posts and even previously generated AI content.  The goal of AI tools is to generate content that is seemingly natural-sounding and written by humans. However, credibility can waver depending on the subject, the sources from which the information is drawn, and the ever-evolving capabilities of the computer program itself. The more information it generates the more it has to draw on.  If there are 20 students all answering the same question for an essay there is a strong possibility that AI would generate 20 very similar pieces of work if asked.  In addition, AI writing can often read awkwardly, with outdated or dialect specific phrasing that can draw unwanted questioning because it sounds out of place.

While AI content might not necessarily fall under the technical definition of plagiarism, is it ethical? There is a strong argument that says no it isn’t. Many academics consider AI writing unethical because students are presenting an idea or text as their own when in fact AI technology created it with very little thought and input from the student.   The student is taking the credit for a piece of work that they haven’t written and generally don’t understand the content of.   Plagiarizing is defined as copying another person’s ideas, words or work and pretending that they are your own. While an AI writer isn’t “another person,” I would say passing AI generated work off as your own still comes dangerously close to plagiarism.  Using AI writing and falsely pretending the content is your own can cause you to get credit you haven’t earned. You may pass your grade with your AI-generated essay, even if you know barely anything about the subject and haven’t put in any work. There are other ethical problems with using AI content. Since AI tools use machine learning and large datasets to get information and create articles, those datasets are likely to be already biased in one direction or another depending on its origins.  Furthermore, AI content does not generate real references for the facts presented, and often such sources used don’t even connect with the content presented.  It can also create false news stories which although based on carefully selected information to support an argument would lead to the presentation of misleading proof.

I would suggest that whatever your topic or subject area don’t use AI to generate your work.  Doing so indicates no understanding of the topic; nor does it help you to expand your own knowledge which would have been gained by reading around your topic and researching what others have said.

What do you think? At Academic Futures we have this discussion with students most days, and especially at the moment with assignment deadlines looming.  We would be interested in your thoughts.  AI is here to stay but how as academics do we manage it more effectively?