AI in action

Artificial intelligence is a useful technology, but it comes with all sorts of risks. How do students use AI tools like ChatGPT? What are the risks of AI? And how good is AI really? In this special, you can read about what is going on about AI at Erasmus University.

AI and students

Do students use ChatGPT for their studies? If so, how? An anonymous survey conducted by EM reveals all the ups and downs of students using the popular language model. “ChatGPT isn’t good enough, and it would be silly because I’m supposed to be learning about subjects myself.”

Through the AI and societal impact lecture series, students explore the impact of artificial intelligence (AI) on various aspects of life. Guest lecturer Claartje ter Hoeven talks about exploitation on the digital labour market.

A year after the release of ChatGPT, students share their opinion on ChatGPT. They address different aspects, such as fact-checking, generating essays, and information access.

Risks of AI

Students voiced their opinion on the use of ChatGPT for their studies, but what do teachers think of it? In a workshop organized by ErasmusX, some of the academic staff discussed the potential and dangers of ChatGPT.

Artificial intelligence comes with a lot of risks, such as discrimination. This is why PhD student Joris Krijger, who researches the ethical aspects of AI, is advocating for a broader perspective on EM TV. “I think it’s dangerous.”

Own AI cloud

Following the developments of ChatGPT usage in academia, Erasmus University will start to make ChatGPT available to students and staff in a protected cloud environment soon. The first users will be able to activate the chatbot early next year.

EM experiments with AI

Can generative artificial intelligence take over the work done by journalists and illustrators? EM experimented with ChatGPT and Midjourney. This article was written entirely by ChatGPT 3.5 based on an automatically generated transcript of an interview about ErasmusX with programme director Vanessa Abel.

While this article does not contain any major factual inaccuracies, there are many issues that should be noted. A description of how EM approached this experiment and a reflection on the results can be found here.

EM had ChatGPT write an article about an interview with the programme director of ErasmusX, Vanessa Abel. This yielded a correct but shallow summary that was lacking in context. For now, it seems, journalistic relevance is still out of reach for artificial intelligence.