Discussions have been heating up about students shamelessly using ChatGPT to write their thesis papers. What should universities do: adapt to the changing world or tilt at windmills?

Goda (name changed, real name known to DELFI), who has recently graduated from university, said that she had used ChatGPT during her studies. According to her, the AI was really useful when writing papers on topics extensively covered online and not requiring a lot of specific theoretical knowledge.

"ChatGPT made summarising quicker and simpler, for example, when writing introductory paragraphs or conclusions," Goda said and added that she had realized right from the start that ChatGPT would not do everything for her. "I used it mostly for topic and subtopic suggestions and similar tasks."

Goda managed to keep her AI assistant a secret from lecturers, but not every student is so lucky. "Even if as much as 60% of the paper is written by ChatGPT, you still paraphrase, change, and add something, so it’s not considered plagiarism, but a friend of mine did not make any alterations to the AI’s version and got caught," she said.

Another student, Vilte (name changed, real name known to DELFI), claimed that she hadn’t heard about students who got caught using ChatGPT.

"I think most of them aren’t that stupid to send their lecturers just a translated text. When using ChatGPT, you still have to paraphrase sentences, improve and edit the text to get your message across. Thus, even if AI tools are being used, usually you can make it work and avoid problems related to plagiarism," she explained.

Vilte added that she had used ChatGPT to improve her paper: "Even though the AI did provide good suggestions, I had to reject some and alter others to better suit my needs. All in all, it was only a utility tool."

According to Vilte, she was introduced to ChatGPT by her fellow students. "I decided to give it a try because I knew that other students had been using ChatGPT and that it helped writing papers, provided ideas, and "put you on the right track". I don’t use it often, though. It is more of a last-resort thing than a robot that does everything for you."

Goda, too, has been influenced by friends.

"I knew nothing about it, while some of my friends had been using it for some time. When they showed it to me, it was pretty new in Lithuania, but already widespread abroad. The AI was like a new inspiration for an easier and faster writing of papers. If the topic is boring, and you find it difficult to feel invested, this platform is the go-to tool is such cases," she noted.

According to Goda, academic life used to be simpler when ChatGPT wasn’t as popular: "It was easier to fool lecturers, but after about six months, towards the end of the semester, more and more people were analysing how the tool worked, what were the peculiarities of its style of writing, and so on." She added that she had to edit the information provided by the AI.

"It would make mistakes, so you just had to read the whole thing. Also, while ChatGPT usually stayed on topic, sometimes it would deviate from it and change the meaning completely," Goda said.

According to Goda, the mistrust of AI exists and will continue to exist, but that’s not the biggest problem she sees.

"Considering the future, I think that the main danger is the fact that people will be less interesting in things, spend less time thinking, putting effort. If this platform remains unchecked and allowed to be used for writing thesis papers, people will become dumber. Why study hard when a program can do everything for you – and faster, better, and easier," she said.

However, although dangers exist, according to Goda, not using such possibilities would be stupid. "If you can get a paper faster and with less hassle, it does not mean you have to copy/paste everything ChatGPT provides. You simply get ideas to make you work easier, to enable developing the topic further."

Goda thinks that cases when students use ChatGPT only as a support tool and do not allow it to write the entire paper should not be considered cheating, but if the AI writes everything, that’s not fair with regard to other students.

"Obviously, the person who actually did all the writing, reading, and research will feel bad," she said.

Vilte also agrees that using ChatGPT as a support measure is not cheating. It’s just a tool that can be replaced with another one.

"Personally, I’ve used it, I don’t think it’s cheating. I would’ve got a very similar idea if I had read some of the works of former students. I guess there’s a line between cheating and not cheating, when it comes to using ChatGPT," she noted.

Vilte believes that it is very important to talk more about the risks, but universities should not ban such AI tools.

"It is imperative to educate students that such software cannot form the basis of a thesis paper and that critical thinking and analysis is required to write it. You are studying for your own benefit, after all. Banning such tools will only distance us from the latest technologies, and lacking the ability to use them may prove disadvantageous in the future," the student insisted.

According to the president of the Lithuanian National Union of Students Paulius Vaitiekus, various reasons have contributed to students turning to AI tools. "Some use them out of curiosity, to verify or find information. Obviously, there are some who use ChatGPT or similar tools to complete curriculum tasks faster and easier, especially those that only require presenting facts, not an interpretation or a personal opinion, argumentation," he said.

However, Vaitiekus also mentioned risks. "Perhaps the greatest risk is the fact that, regardless of how convenient and simple it is to use AI tools, the student cannot be completely sure that the information is accurate, and the sources – reliable. This poses the risk of having poor end results. Also, lecturers may notice the use of an AI tool and accuse the student of plagiarism," he added.

The president also mentioned that lecturers were becoming more vigilant due to ChatGPT’s increase in popularity.

"It does not really matter whether the student plagiarizes somebody else’s work or uses AI tools, the consequences are similar: from a failing grade to expulsion. Nevertheless, in academic ethics the use of AI tools is somewhat of a grey area since there are no universally agreed upon rules and regulations how to assess such cases," Vaitiekus said.

He continued that it would be naïve to think that students would not use technologies as they got better, but certain boundaries should be established. "Clear lines need to be drawn on what is considered cheating. We believe that the most important thing is to ensure that students take responsibility for the information they submit, its accuracy. AI can be used as an information generation or search tool, but the author should be the student," Vaitiekus said.

According to him, higher education institutions should take into consideration the fact that the world is constantly getting more modern, thus the possibilities that AI provides should be incorporated into the studying process. But that should not mean tolerating cheating or plagiarism.

"Ethical behaviour in the academia is one of the fundamental principles of higher education and it has to be upheld. On the other hand, the relevance of tests and research papers spanning several dozens of pages that seemingly nobody actually reads should be re-evaluated. The student’s ability to think, interpret, discuss, create, etc. should show their knowledge and skills, and in such situations it would be difficult to just blindly copy everything," he insisted.

According to Kaunas University of Technology pro-rector Jonas Ceponis, both students and lecturers, just like all the other people who want to try new technologies, are using ChatGPT and similar programs.

"It is a tool that can provide answers on various topics. But it is vital to check how students understand the generated text, assess it critically, and are able to interpret," he said.

However, Ceponis stressed that AI tools may have a negative effect on our critical thinking.

"One may develop too much confidence in generated answers, which may not necessarily be correct. Also it is worth mentioning the weakening of practical skills development due to the use of pre-generated texts or software codes," he added.

Source: https://www.delfi.lt/multimedija/mano-poveikis/universitetu-laukia-dideli-isbandymai-studentai-rado-kam-patiketi-darba-kuri-turetu-atlikti-patys.d?id=94034355#article