By Kaya Bogot

The end of 2022 surprised the world with the launch of ChatGPT, a new artificial intelligence (AI) chatbot. OpenAI launched the program as college semesters were ending and finals were beginning.  

The timing of the program’s release was one that may have affected college finals. As many students dealt with the stress of finals, some may have turned to shortcuts and cheating to relieve stress over assignments, ChatGPT was an accessible solution.

ChatGPT is available online for free, making it easily accessible for anyone to use it. The chatbot can converse with users asking for specific information and allows for follow-up questions. 

The AI’s algorithm reviews billions of pieces of information from across the internet to provide a response to whatever the user is asking of it. For example, this program has sparked a lot of conversation online, in the news and within academic communities. This is what ChatGPT had to say about this issue:

The program also knows every language, including coding languages, giving it the unique ability to write and correct code.

While this new AI is a step forward in computer science, many people question the future of writing and coding, and the threats AI poses, specifically to education.

Plagiarism

With ChatGPT, the temptation and threat of plagiarism may be on the rise across schools and universities. With its ability to write detailed responses to a prompt and code based on simple instructions, plagiarism could be easier than ever. 

Here’s an example:

“I think it is plagiarism,” Ming Lin, a computer science professor said. “How do you define plagiarism? Taking someone else’s work. That’s what an AI is doing, and if you are using an AI to generate your homework for you, isn’t that plagiarism? Isn’t that cheating?”

Additionally,  artificial intelligence makes it more difficult to spot plagiarism. 

Educators are also worried about students’ ability to learn through practice with the potential temptation of ChatGPT. Ultimately, the students decide to cheat or to learn. 

“When students are taking classes, you can argue their goal is to get good grades, but I would argue that they are coming to college and going to classes to learn for life,” Lin said. “But when they plagiarize someone else’s writing through AI technology, you bypass that process, you bypass that learning process and you don’t learn.

However, in the past few weeks, OpenAI announced it will launch a program capable of detecting the likelihood that something was written by AI called “AI Text Classifier.” However, this program has limitations:

• “Requires a minimum of 1,000 characters, which is approximately 150 – 250 words.”
• “The classifier isn’t always accurate; it can mislabel both AI-generated and human-written text.”
• “AI-generated text can be edited easily to evade the classifier.”
• “The classifier is likely to get things wrong on text written by children and on text not in English, because it was primarily trained on English content written by adults.”

This is according to OpenAI’s website.

Here’s an example of ChatGPT responding to a writing prompt:

How should this issue be dealt with?

For many schools and universities, this program presents a new challenge. James Bond, University of Maryland’s student conduct director, advises that instructors should have a candid conversation with their students about cheating.  

“If you don’t want ChatGPT to be used in your classroom, then make sure that is clear to your students both in writing, and speak about it when it comes to specific assignments,” Bond said. “Some instructors may want to talk about ChatGPT and how it could be used in the classroom.”

Maryland’s Office of Student Conduct will look into cases where students may be using ChatGPT on an assignment. “If you trust the students’ work was done correctly, that’s fine, but if you have a suspicion, let’s look into it. And then if it’s referred to our office, we will talk with the student and learn about how they went about completing the assignment,” Bond said. “Trust but verify.”

Bond also says if a cheating issue concerning ChatGPT is brought to his office, it will be dealt with in the same way as any other cheating case. This involves talking to the instructor to find out what their expectations were for the assignment and talking to the student to learn how they completed it.

Benefits

Despite the risks, professors like Lin think a collaborative version of ChatGPT can benefit teaching.  The chatbot could be used to generate exam questions and assist with grading. In computer science, ChatGPT can correct code, grade assignments and offer an opportunity to show students where they went wrong.

“If we have a program where the student can have such a tool, maybe offered through the instructor or the classes to practice programming at home, to have it identify mistakes, it can help a lot,” Lin said. 

It’s not perfect

Even if a student were to use this tool to cheat, it doesn’t always guarantee an A+ assignment. One of the biggest flaws with the program is its accuracy. ChatGPT’s comprehensible responses are not always factually correct, which can generate or disseminate misinformation. ChatGPT has even generated fake sources in papers, according to NBC. 

Ultimately, some professors aren’t worried about ChatGPT just yet. 

“Because large language models don’t actually understand anything they output, they are, at root, mimicry machines–  they can imitate, but they can’t innovate. I guess that’s where we still come in,” Matthew Kirschenbaum, an English and Digital Studies professor at UMD, said. 

“It’s very surface level, it has no soul,” Lin said. 

The Answer

As AI technology continues to develop, educators and students may need to navigate the digital space with caution. 

“I think it has the potential for being harmful but I think being overly cautious is good right now, and if we are overly cautious and if it doesn’t end up being as big a concern down the road, great! But let’s air on the side of caution, learn as much as we can about it,” Bond said.

Whether or not a student decides to use the service to cheat is their decision, but the consequences for cheating with ChatGPT will likely be the same as any other form of cheating. 

Since ChatGPT is easily accessible to students, the dangers of cheating could become more prevalent. 

Ultimately, schools will need to adapt to the presence of this technology because these programs are not going away as more companies like Google and Microsoft invest in AI technologies.

“The more I learn about it, the more I am going to share with others about it and how we can work within our policies and then also work within our instruction, with instructors, and students about its use.” Bond added.