AI is more than just ChatGPT

How students use AI in their studies and everyday life

Context

The Teaching Day is an annual gathering of professors at Offenburg University of Applied Sciences. Students can attend to help shape teaching conceptually. This year, I participated as a representative of my degree program “Enterprise and IT Security” with this presentation.

The presentation was given again shortly afterwards at the KI-Ideenzirkel Ortenau.

Content of the Presentation

At the time of this presentation, AI was becoming increasingly popular. The professors at Offenburg University of Applied Sciences feared that teaching might suffer as a result. Equality with students who cannot afford an AI subscription was also part of the discussion.

In this presentation, I anonymously surveyed my fellow students to find out which AI systems they use for what purpose. This was intended to show the university’s teaching staff that AI is more than just ChatGPT and “I’ll quickly send my math exam to the chat and have it solved.” The following general insights were gained:

  • 60% of all participants have searched for AIs that, unlike ChatGPT, are aimed at specific applications
  • Everyone who used AI for homework or code generation stated that they always manually checked generated content
  • Students never assume that AI answers are completely correct – some assume “gross negligence” in responses
  • AI hallucinations, i.e., the generation of obviously false statements, were recognized as such by 70% of participants
  • All participants define prompting as a necessary skill in today’s world
  • About half of the participants have strong data protection concerns and privately deal with locally operated AI
  • 85% of participants believe that students should be able to decide for themselves whether AI may be used
  • 15%, on the other hand, fear a diminished learning effect
  • Consensus: The possibilities of AI in a university institution MUST be addressed – a strict ban is unanimously rejected

The participants named the following areas of application; in ascending order of frequency:

  • Summarization/structuring of scientific documents
  • Writing/rephrasing emails
  • Execution of repetitive tasks in programming
  • Automated creation of flashcards
  • Creation of practice exams
  • Writing/rephrasing homework

One student gave an interesting example of the use of AI in a scientific context, which I also used as a closing word in the presentation:

You can imagine working with ChatGPT to write a homework paper as follows: You sit at a table with 10 people and distribute research work to them. After 2 days, everyone comes back together and presents their results. Some of them did their job well, others did not. As a project manager, you now have to find a consensus. That doesn’t work without your own knowledge and competence.

Replace the 10 people and the 2 days in the example with ChatGPT and 2 hours, and you have a term for the way of working and the responsibility you have when working with it. Unreflective work is dangerous even without AI. You must not let yourself be dazzled and must remain critical, but the opportunities are enormous.

This presentation clearly shows that students, despite the great dangers of AI, deal with it critically and are very aware of the dangers.

References