Keeping up with the AI revolution is a difficult – perhaps impossible – task for universities. Students and staff are quickly embracing the technology as a means to simplify their work and studies. Academic institutions can’t ignore that. But the risks pertaining to privacy, safety, sustainability and the quality of education cannot be ignored either.
These risks are likely why the general policy of VU Amsterdam on AI remains that it should not be used unless a lecturer allows it. This means that some students barely use AI in an official fashion (unofficially they may be using it regardless), while for other students it has become a widely used tool.
Judging by the many articles and updates about AI on the VU website, the university does take AI seriously. Still, there is much left to be decided and arranged.
Master’s student Cognitive Neuropsychology Isabella Ivory says that it’s great that there is a lot happening policy-wise, but she worries that the sense of urgency may be missing somewhat. “We’re a bit late already and there’s still not a clear policy and still so many teachers who absolutely have no clue and are not making their curriculum AI proof.”
Copilot subscription
VU Amsterdam does offer an AI tool to its students and staff since February 2024, namely Microsoft Copilot. If this is news to you, you’re not alone. Isabella Ivory of the Education Lab says that she often hears from teachers that they never knew that they had a work account for Copilot. And while this AI tool may be safer than the larger Copilot 365 version or other tools, Ivory says it gets less use because it’s not up to par.
Learning from influencers
Third-year Artificial Intelligence student Natálie Halašková says she mainly uses AI to explain concepts that are difficult to digest. “If you prompt ChatGPT to explain university materials as if you are fifteen years old, it will do so in an understandable way.”
As a student of Artificial Intelligence, you might expect that Halašková receives a lot of guidance on using AI, but that is not the case. “So far we got a lot of theory: how things work. But we don’t get information about how people use it.” Instead, she says it’s mostly up to the students themselves to figure out the practical part of the equation. “All my friends get this information from Instagram or from TikTok, where you see influencers using AI”, Halašková explains.
This aligns with the experience of first-year AI student Jasper Meijerink, who says he too gets his information about the practical use of AI from social media. He says his lecturers don’t really teach about how to use AI tools specifically, except for things like grammar.
According to Halašková, students are told they can use AI, but receive no guidance on which tools to use. They are thrown in at the deep end and have to figure their own way to swim out – a teaching method she says is common at VU Amsterdam. One of the topics she feels is missing is benchmarking: a way to compare how different AI models perform in different categories. “That is basic and a lot of people who are not really into AI still don’t know about it.”
She and her friends think that practical use of AI is the most important skill in the job market. And considering that maybe less than ten percent of them will end up in academia, such experience could be very valuable to a lot of students.
Student becomes teacher
Of course it’s not just students who require guidance with the use of AI, but teachers as well. Perhaps even more so. Hence the VU Education Lab gives (voluntary) workshops to teachers, some of which are hosted by student assistants such as Meijerink and Ivory.
“There seems to be this dissonance where students know more than professors when it comes to the effective usage of chatbots, how to prompt better or what the possibilities are”, explains student assistant and Computer Science student Amelia Kalecińska. She says that the workings of AI can still appear like “black magic” for some.
Kalecińska thinks the reason for the difference may have to do with student’s open and experimental attitude towards AI. “I noticed that my parents are a bit scared to experiment with AI on their own.” Meanwhile students learn through trial and error, from each other and from social media.
Esther Schagen of the VU Centre for Teaching & Learning adds that while students may be more skilful and better at searching with AI, teachers are better at other aspects of AI such as validating the output.
Isabella Ivory, who gives AI workshops, says that it’s great that social media makes AI more accessible. However, she worries that influencers can present AI as a “cool new tool”, while ethical aspects get left out. “There’s probably many students who are uploading their entire bachelor’s thesis on there and I don’t think that’s so great.”
Wasteful usage of energy is another problem according to Ivory, with people asking random sarcastic questions with no real purpose, or using AI as a replacement for Google. “Since VU is like the most sustainable university, you would think they would do a little more about it”, she says.
Not stuck
AIKI hosts monthly events about things such as AI in health, AI in law and AI in finance. Besides that, the club has bi-weekly recommendations for tools – giving students alternatives to solely using ChatGPT – as well as weekly news updates to explain recent developments in AI. The club is free and can be joined through their LinkedIn and Instagram channels.
Natálie Halašková sees a solution to improving students’ understanding of AI in the creation of AI-related communities. During her exchange in Singapore, such initiatives were being pushed by her university as well as the government. She hopes to bring the collaboration and spirit she saw there to VU by founding the association AIKI. Those last two letters stand for the Dutch translation of AI: kunstmatige intelligentie.
AIKI was founded earlier this year, although it is not an officially recognized student organization yet. In February it held its first event, welcoming quantum professor Paul Griffin as a guest speaker to introduce machine learning.
While Natálie Halašková thinks the university is dropping the ball on some aspects of AI, she does notice that things are changing every year. That means VU Amsterdam isn’t necessarily stuck in its ways. How long it will take to truly get a handle on AI is hard to say, but programme manager at the VU Education Lab Silvester Draaijer says that there are some major developments underway: “In the context of the institutional plan, VU Amsterdam will boost AI literacy for both students and staff this year. Faculties and services will work together to formulate final attainment levels and to develop curricula and teaching materials to integrate into VU study programmes.”