ChatGPT Makes It Easier to Plagiarize in Schools, Study Warns

A smartphone with a displayed ChatGPT logo is placed on a computer motherboard in this illustration taken on Feb. 23, 2023. (Dado Ruvic/Reuters)

The rollout and subsequent popularity of artificial intelligence (AI) chatbot ChatGPT has brought with it significant stresses in the education sector, with concerns about academic honesty and plagiarism brought to the forefront following the adoption of such technology.

Three researchers from the University of Plymouth and Plymouth Marjon University in the United Kingdom set out to demonstrate this in a paper published in Innovations in Education and Teaching International.

“One challenge with using GPT-3 for assessment in higher education is the possibility of plagiarism,” the paper said. “AI essay-writing systems are designed to generate essays based on a set of parameters or prompts. This means that students could potentially use these systems to cheat on their assignments by submitting essays that are not their own work.”

Academic staff and exam evaluators find it difficult to distinguish between answers derived by the student and those produced by a chatbot.

“This undermines the very purpose of higher education, which is to challenge and educate students, and could ultimately lead to a devaluation of degrees,” the paper said.

But near the end of the paper in the “discussion” section, the researchers—Debby R. E. Cotton of Marjon University and Peter A. Cotton and J. Reuben Shipway of Plymouth University—revealed that everything preceding the discussion section, including the excerpts highlighted above, had been generated using ChatGPT.

“Our intent in taking this approach is to highlight how sophisticated Large Language Machines (LLMs) such as ChatGPT have become,” the authors wrote.

Preventing Students From Plagiarizing Using ChatGPT

The researchers fed several prompts to the chatbot to generate the resulting paper, such as:

• How can university staff design assessments to prevent or minimize the use of GPT-3 by students?
• How can academics prevent students plagiarizing using GPT-3?
• What are the challenges of GPT-3 for assessment in higher education?
• Write an original academic paper, with references, describing the implications of GPT-3 for assessment in higher education.

The resulting paper encourages academic staff to adopt a combination of automated and manual assessment techniques instead of standardized testing approaches. Besides educating students on plagiarism and compelling honesty through written declarations, staff should be required to employ AI-detection tools, the paper says.

“For example, they may use natural language processing algorithms to analyze the language and style of submitted work and identify any anomalies that may indicate the use of ChatGPT or other AI language models,” it says.

In order to detect AI work, the instructors can look for patterns and irregularities in the language as chatbots currently have limited language abilities, and their method of output suggests a nonhuman quality. Chatbots cannot be relied on to generate original, unique and innovative content.

Hence, checking for originality should be part of test assessment.

A student is typically not flawless, while AI tech strives to be so. Therefore, when submissions come in with perfect grammar and spelling, the paper can be evaluated further. Human writing is generally more aware of its audience, while the AI, for now, sticks to the context of the subject matter, the paper says.

Spread of AI

Developed by AI research company OpenAI, ChatGPT is a natural language processing tool—a chatbot.

Since its public release in November 2022, the tool has garnered hundreds of millions of users in record time initiating an AI race. The latest version of the language model system is GPT-4 with free and premium versions.

OpenAI had over 1.1 billion users visit its website in February 2023, an increase of over 60 percent from January’s 667 million, indicating its expanding popularity across the world. Most of its users are young, between 25 and 34 years old, according to Similarweb.

As with all technology, ChatGPT comes with its own limitations in intelligence and privacy. A data breach was recently discovered, and the platform was taken offline for maintenance on March 20 as users reported seeing chat histories from other users.

The bug has since been patched, with CEO Sam Altman issuing apologies for the exposure of private information, including payment details.

ChatGPT has also been accused of left-wing bias in many of its answers. The technology is still in its nascent stage in the public sphere, and it remains to be seen how AI will evolve.