This article has been reviewed according to Science X's editorial process and policies. Editors have highlighted the following attributes while ensuring the content's credibility: ChatGPT hit like a meteorite in late 2022 and found fertile ground at EPFL, invading the computers and smartphones of tomorrow's engineers. Within two years, generative AI had spread like wildfire among our students owing to their still-malleable learning habits and the growing number of models—and the many improvements to them.
The good news today is that AI has replaced neither our students' brains nor their memory. While over 90% of EPFL students use it for their coursework, they do so mainly as a learning tool. For them, it's a quick and easy way to lighten their workload and expedite large and small tasks such as scheduling their study time, taking notes, completing class assignments, learning from their mistakes, taking practice exams and managing stress.
Around three-fourths of these users find it effective for explaining the course material, and two-thirds for helping them solve their weekly homework problems. How students use AI also depends on their major. Those in computer science, for instance, take a more advanced approach, while those in mathematics are more wary of the models' statistical predictions.
Those are just some of the key findings of a survey carried out at EPFL by an AI working group made up of the school's student delegates. The group sent the survey to around 12,000 bachelor's and master's students last winter and received 359 replies from a representative sample of the different majors and levels of study. One finding that some may find reassuring is that 40% of respondents don't use AI outside of their coursework.
And of the 60% who do, it's mainly "to get answers to questions that come up during the day." However, just because students claim to use AI as a learning tool doesn't mean they do so correctly. To help students make the most of the technology, EPFL's Teaching Support Center (CAPE) published two handbooks this year. They serve as a compass for students, asking them whether AI is helping them learn or is doing the learning for them.
"Our main concern is that students are delegating cognitive tasks, or in other words outsourcing some of the mental effort to a program or machine," says Olga Kasatkina, a researcher and teaching adviser at CAPE. "We don't yet have enough experience with AI to determine how it's affecting the learning process." The survey revealed that students generally have a fairly objective view of AI's pluses and minuses. They're aware of its strengths and its potential as a powerful tool when used properly, and they're worried about its biases, the scope for misuse (such as to polarize debates, infringe upon privacy or breach IT security) and its damaging social and environmental impacts.
Extract — continue reading at the source.
