‘I see ChatGPT being used everywhere and by everyone’, says a bachelor’s student in anthropology. Like many students, she finds herself relying increasingly on the AI bot, which is developing at breakneck speed and is embraced by students all around, though feared by lecturers.
‘I can hardly imagine a student life without it,’ says a master’s student in Asian studies, who, like the other interviewed students, only wants to speak to Mare anonymously, out of fear of being accused of plagiarism. ‘I find it difficult to write a thesis without ChatGPT. It’s really chill when ChatGPT can rephrase sentences for me. For example, I ask: “Can you please explain this to me like I’m five?”
‘For my bachelor thesis, I mainly used it as a brainstorming tool: by asking ChatGPT questions, I eventually arrived at my thesis topic and research methods. I also had to read and compare various articles. Using ChatGPT, I was able to do this much better and faster. It's good at summarising as well. So yes, I do feel quite dependent on ChatGPT... quite troubling really!’
SIMPLY FASTER
‘In our programme, there's a lot of reading, sometimes hundreds of pages for a single lecture, says the anthropology student. ‘Those are the moments I use ChatGPT. I could do it myself, of course, but with ChatGPT, I’m simply faster. I don’t necessarily feel dependent on it, but it affects my life much more than I care to admit.’
The chatbot is also popular among science students, says a bachelor's student in physics. ‘So many people use it. In the library, I often see open tabs with ChatGPT. Students literally have lists of questions next to their computers and ask AI to solve equations. They do adjust the brightness of their laptop screen to make it less obvious.’
She also uses the bot herself. ‘In the first year, you learn programming. In the second year, you’re supposed to apply it during lab hours. During these lab hours, I secretly use ChatGPT when the lecturer isn’t looking. I have it program codes, and then my lab buddy and I try to figure out what it generated; we then check it by calculating and use the answer. My standard question to ChatGPT is: “This is my question, can you please code this for me?” Then I memorise the steps it generates. So basically, I learn from the chatbot instead of my lecturer.’
However, this does not always work. ‘At the end of our programme, we have to do a bachelor research project. This project requires you to write codes and process data in a lab while someone watches over your shoulder. A student who had previously received an eight for programming thanks to ChatGPT soon realised: I can’t do this. He quit and is now studying mathematics.’
PAID VERSION IS BETTER
Another master’s student says she recently bought the paid version, for 22 euros a month. ‘I was on a wait list for a month. That goes to show how high in demand it is.’ According to her, the paid version is much better. ‘It gives you access to more literature and can also create PowerPoints.’
Although the chatbot seems to be widely adopted among students, lecturers are divided on ChatGPT and the way in which the university responds to the technology. As of last year, a few brief rules can be found on the university website, which essentially boil down to: using it as a sparring partner is allowed, having it write a text is not (see text box).
‘A lot of time has been wasted,’ says Tom Theuns, assistant professor of political science, a member of the Programme Committee and chair of the AI taskforce. ‘It took the university a very long time to arrive at a policy, even though we saw it coming. This caused a lot of frustration among lecturers.’ Given the rapid developments, ‘the regulations should also be changing promptly,’ says Theuns. ‘But they’re not.’
According to university spokesperson Caroline van Overbeeke, it is still ‘too early’ for further adjustments. ‘In February, the project group “AI in education” was launched, which will most likely lead to policies on the use of AI in our education. This group brings together experts from across the university, with both educational and substantive expertise. Students, lecturers and examination boards are also involved. We can’t provide much more concrete information at this point.’
FROM PANIC TO UNDERSTANDING
In the meantime, the faculties are trying to adapt their policies themselves, including Social and Behavioural Sciences. ‘A content tab is now created automatically on all Brightspace course pages on the use of AI,’ says Cynthia van Vonno, chair of the Bachelor’s Programme Committee for Political Science. ‘Students also have permission to cite its use.’ A small note: ‘It’s up to lecturers to allow it or not, and they need to be very explicit about what is and isn’t allowed. That also has to be explicitly stated in the papers students submit.’
The fact that policy is always lagging behind is natural, Bas Haring believes. ‘Patience is advised when it comes to these kinds of things,’ says the professor of public understanding of science, who teaches the ‘AI & Ethics’ course. ‘Not allowing the use of ChatGPT is very foolish. It's here, and it’s here to stay. We need to be open and honest about its use. You should ask yourself: do I want to learn or do I want a high grade?’
‘When ChatGPT was first introduced, the university was caught off guard,’ says Wim van Anrooij, chair of the Dutch Studies Board of Examiners. ‘The initial reaction was panic and lecturers were mostly of the same opinion: it shouldn’t be allowed. Now there is a lot of understanding. They think: it’s here now and we have to deal with it.’
To gain further insights, he started experimenting with it himself. ‘I asked ChatGPT to write me a poem in the style of Louis Couperus. What rolled out was an awful poem, but still, I recognised certain elements. For example, Couperus used gems in his poems, and ChatGPT did the same thing.’
CONTROVERSIAL QUESTIONS
Theuns has also been trying out the chatbot. ‘A lot, because I want to have a good understanding of how it works. Personally, I use it as a kind of turbo spell checker, for example. It’s good at creating standard texts for which there are already thousands of versions available. But when you want it to do something creative or original, the results are disappointing. It’s also not useful for conducting research. And it doesn’t do a very good job providing exact information. For example, it knew who the first female mayor of the Netherlands was, but not the first female mayor of the province of Overijssel. Furthermore, the bot often refuses to answer controversial yet scientific questions.’
‘My biggest concern is that students are gradually losing sight of the fact that its use is wrong’, says Van Anrooij. ‘It's not about: will I get caught? It’s just like smoking: you know it’s bad for you. ChatGPT is a programme that automatically generates answers. If you use it, you’re no longer thinking for yourself.’
Theuns: ‘I don't think using ChatGPT as a writing aid is cheating. But if you use it to find an answer to a research question, it becomes much more problematic. However, when it comes to fraud, proof is important, and that’s difficult to come by in this context.’
‘Proving that somebody has used AI is virtually impossible,’ says Van Vonno. After all, we don’t monitor students as they’re writing their papers. We can only catch students in the sense of them admitting to using AI. And as far as I know, there are only a few cases where that has happened.’
That is why the university has to learn to deal with it, says Theuns. Testing methods need to change. Think of keeping a log during research or defending your thesis orally, those could be possible changes.’
LAZIER
‘You'd have to be very thick to have it write your entire essay’, thinks the master’s student in Asian studies. ‘I would never do that. Integrity is important to me. I want to be proud of my master's degree.’
‘It worries me that AI has made its way into education’, says a master’s student in Middle Eastern studies who has never used ChatGPT. ‘All it does is make students lazier. I know a lot of people who don’t do the reading any more, but simply run the PDF through ChatGPT and only read the summary. Of course it’s easier to get your answers from ChatGPT and directly copy them, but I don’t think you learn anything from it. You lose certain skills, like summarising, critical thinking and academic writing.’
What is and isn’t allowed when using the chatbot? ‘Students can use ChatGPT as a sparring partner’, it says on the university website. ‘This means that ChatGPT can be a useful tool during the beginning stages of your research. For example, it can help you whilst brainstorming and it can help you search for information. Be aware that ChatGPT does not always give you factual information. This means that you have to check the validity of its replies yourself. ChatGPT can also help you structure your text.’
What is not allowed is ‘to let ChatGPT write your text, or to let it rewrite an already existing text’, it says on the website. ‘This is seen as plagiarism, as your text will be based on someone else’s words and the original source will not be cited. ChatGPT is good with tone and structure, but lacks interpretative skills and cannot understand the text it processes or generates. Sometimes the tool also fabricates references and sources. These are not usable for your research.’
The website also includes a brief side note. ‘Chatbots are developing rapidly. This means the rules and regulations surrounding them must be updated on a regular basis. Always check the Leiden University website for more information concerning the use of ChatGPT.’