The ABC of Artificial Intelligence

Machines that can perform tasks that used to require humans – for decades, this was the stuff of both utopian and dystopian fantasy. In the space of just a few years, the rapid advancement of artificial intelligence (AI) has made it part of everyday life – including at university.

This immediately raises some practical questions: what tasks do we want to leave to machines, and which ones don’t we? How can we tell whether an output is right if it sounds plausible? AI literacy is now as much part of tertiary education as statistics and academic writing.

The Faculty of Science therefore developed a course in a short space of time and integrated it in all study programs. Around 1,100 students have already completed the module called AI111 Artificial Intelligence and Critical Thinking. The technical foundations of AI are just one part: students learn the fundamentals of how a language model works – and why it isn’t conscious, even if it might seem like it is when you chat with it. They practice structuring prompts and systematically checking answers in terms of completeness, factual accuracy and biases. They discuss the energy consumption of these tools and the danger of skill atrophy, where people risk losing the skills they no longer practice. 

But the real starting point comes before that: what is the essence of scientific method, and how has it changed over time – up until the recent AI boom? Where is the limit between an exploratory and a confirmatory experiment, and what happens if a tool blurs this boundary? The aim is to develop what the course calls epistemic independence – the ability to assess what we know and how we know it.

Three days of lectures are just the start. In the subsequent semesters, subject-specific courses will be added in which students apply what they’ve learned in the relevant discipline. The early timing of the course is intentional so that students complete it right at the beginning before habits take hold. 

How did this early start to the semester go down with students? At first, she wondered about the timing of the course, before the actual start of the semester, explains biochemistry student Mia Riina Lampinen. “But in retrospect it was a good way to start, and at the same time it gave me the opportunity to get to know the Irchel Campus a bit better.” Benjamin Furrer, who is starting a degree in physics and Latin, welcomes the fact that all new students start their studies from a shared basis. “It’s good that everyone is on the same page regarding this important topic.”

The two new students already used AI extensively at school – for homework, summaries, explanations and to structure work. They also became aware of its downsides: “For difficult tasks in particular, you learn most when you try and figure it out yourself. If you outsource your thinking to AI, no learning happens,” says Lampinen. Furrer has a similar view. He says that it is particularly tempting to use AI in areas where you’re struggling. But that also poses a risk: “If you let AI do everything, you lose your autonomy and basic skills.”

The course addresses topics such as how machine learning works, the role AI can play in analyzing large data sets, and the methodological and ethical questions raised by its use in research. “AI is a powerful tool, and it’s incredible the number of things it can be used for in research,” says Benjamin Furrer. “But it’s important to start by being clear on the scientific question and the suitable methods – and only then to decide whether and how AI should be used.” 

AI can be a very useful tool in science, he says – provided we understand its limitations and can assess its output. “We have to remain critical, think for ourselves and only use AI intentionally where it makes sense to do so,” says Riina Lampinen. This is the key takeaway from the course for both new students.