AI is here to stay at UvA: “You can't escape it”
Is the UvA integrating applications of artificial intelligence into education? Not everyone is open to it right away, says Associate Professor Erwin van Vliet. “But when success stories emerge, most are very curious.”
“I think Artificial Intelligence (AI) is the biggest revolution since the industrial revolution,” Bart van den Broek dares to state with certainty. But he has a certain right to speak. His field of study, Information Science at the University of Amsterdam, partly concerns that topic: the use of AI by humans and its impact.
But using it himself, for his studies or work? “I still find it exciting.” Agreed, the cover of his thesis was created with Dall E, a program that generates AI images. “But I don't quite trust the output of a program like ChatGPT yet.”
The same applies to econometrics student Gideon van Diggelen (24). He too (still) prefers the old, familiar search engines to generative AI tools like ChatGPT or Google's Gemini. “I like them better.”
Still, both agree: AI is here to stay at the university. Not only is the popularity of studying artificial intelligence increasing structurally, more and more students and faculty are using AI tools. You can't escape it, is the sentiment, whether you like it or not.
How is the UvA dealing with this? Is the university embracing applications of AI and enabling students and faculty to experiment with AI tools? And if so, are they catching on at all? Or does the university prefer to stay at a safe distance, without integrating artificial intelligence programs into education?
Learning modules
The former, it seems. Recently, UvA's Teaching and Learning Center (TLC) introduced a module that teaches students how to responsibly handle generative AI (GenAI). “We know that students use ChatGPT,” explains Erwin van Vliet, associate professor and program director of the undergraduate psychobiology program. “And that can also lead to abuse, which is why we want to educate them.”
“Transparency is essential,” according to the e-module, which takes about three-quarters of an hour. Then the use of GenAI is possible, provided it is expressly authorized by a teacher and properly referenced to sources. The UvA's fraud and plagiarism rules apply unconditionally.
When Van den Broek began his thesis project, his lecturer told him: “In principle, you can write your entire thesis with ChatGPT, as long as you can substantiate why that is a good choice and are transparent.” Van den Broek says: “Basically, lecturers know it's a tool that's going to be used and they can't stop it.”
Personal assistant
Besides, you can also use it for editing, search assistance, brainstorming, and studying for exams, according to a Humanities faculty guideline called “Generative Artificial Intelligence in Education.” In these ways, an AI tool can certainly be a useful addition, notes psychology student Sam (23). He says he uses it as a personal assistant and as a means of coming up with ideas.
UvA lecturers can also benefit, of course. So last week, they also received an e-module from the Teaching and Learning Center in which they learn how to responsibly use generative AI in their
work. And that's not all, as evidenced by several pilots launched at the UvA since last March. Because, in Van Vliet's words, AI is more than large language models like ChatGPT.
One of those pilots concerns automated, personalized feedback on completed assignments by students. Van Vliet is the project leader of that. The pressing question to him is: Does that go together, feedback that is both automatically generated and personalized?
Feedback
“Yes, definitely.” He illustrates how it works. Our own tool—called IguideME—gets access to the digital learning environment and, for example, to answers for practice tests, so it knows exactly on which parts and topics students are missing points. Then, in the form of personalized feedback, the tool can point students to parts they need to pay extra attention to.
Psychobiology student Jazz Stofberg took the test. His experience was positive, especially because he could compare his study progress with other, anonymized students. “Quite a nice approach,” he says. If he was aiming for a seven, he could see how much time other psychobiology students who were also aiming for a seven spent on the course each week. If he noticed that he had spent relatively more time on the course that week, he would leave it at that.
Of course, ideally you would have someone monitoring each student's practice assignments and giving personal feedback on where someone is in the study process, Van Vliet acknowledges. “But we simply don't have enough teachers for that and too many students. A teacher of a large program may well get to know some students, but certainly not all eight hundred.”
Not always true: GenAI gives plausible answers that are not necessarily correct.
Biased: GenAI uses online databases and texts that may contain biases.
Obsolete: ChatGPT uses only available information that has been online until January 2022.
Source reference: ChatGPT does not reference sources.
Privacy: It is not clear what ChatGPT does with personal data.
Keep in mind at all times that ChatGPT and other GenAI tools are owned by private companies.
Inequality: A paid version of ChatGPT provides more features than an unpaid version.
Skepticism
What do teachers themselves think of these new developments? Don't they see important parts of their beloved profession being unwittingly taken out of their hands by these new, technological developments?
“Not everyone is immediately open to it,” Van Vliet acknowledges. Certainly not the more conventional teachers. At first, they say: None of this is necessary for me. But what I notice in research and education is that as soon as a small group of teachers uses a tool and is enthusiastic about it, it spreads like an oil slick over the rest.”
Take Perusall, an AI tool where students can read and annotate a text together. “We first rolled that program out on a small scale in my course. Then several colleagues took to it, after which it expanded throughout the UvA,” even though there was initially a lot of skepticism about it. “But when success stories emerge about AI applications, most people are very curious.”
