• 9 oktober
  • Talentgericht

How to use AI in education? Erdinç Saçan and Colette Cuijpers collaborated on brochure

Colette Cuijpers and Erdinç Saçan.
Fontys has had a workgroup artificial intelligence (AI) since January. They have already produced two versions of a brochure this year, drawing employees' attention to the benefits but also the pitfalls of AI. No hard rules, but manuals, according to two employees involved, lecturer-researcher Erdinç Saçan and lector Colette Cuijpers.

'Educational institutions around the world are currently considering how to deal with AI,' says Erdinç Saçan, lecturer-researcher at Fontys ICT. 'We want to be able to use AI responsibly, while maintaining our academic integrity.'

Ignoring or banning AI is not an option; industry and students are already using it. 'That is why we are in favour of teachers experimenting with AI, but at the same time being aware of the dangers. Hence we have prepared a brochure with guidelines. No hard rules but handouts or suggestions.'

Giving clarity

Colette Cuijpers, lector at Avans-Fontys Law University of Applied Science, believes Fontys should provide clarity to both teachers and students about what is allowed and what is not. 'AI is a tool that potentially offers many advantages but can also be detrimental to education. We want to exploit those advantages, but we need to avoid those disadvantages. We don't think strict control or even prohibition, as has been suggested, is a good idea. Of course, we too do not want students to use AI to obtain a degree they would not have obtained otherwise.'

'That's why we need to start focusing more on our learning objectives. If a writing assignment is your learning objective, I wish you good luck. With ChatGPT, you can easily get a reasonably persuasive piece of text written and it is almost impossible to prove that it was generated by AI. Perhaps the learning objective will be: write a good prompt - a well-worded query that gives a high probability of a good search result - for ChatGPT.'

Erdinç confirms that. 'Writing good prompts is expected to become an important task. There are already names for that function: prompt engineer or AI whisperer. Writing a good prompt is essential for a good answer, because garbage in is garbage out.'

In terms of copyright, the text of the brochure remains fairly vague, Colette admits. 'If something is copyrighted, you are not allowed to just put it online or enter it into a system like ChatGPT. We deliberately stayed vague because all over the world there are currently lawsuits against AI for alleged copyright infringement.'

'This means that, as a teacher, you can't just let ChatGPT assess all your students' papers either. We deliberately did not want to be prescriptive. However, we do believe that, as a lecturer, you should set the rules for what is and what is not allowed with AI within your subject,' says the lecturer.

Limited shelf life

The shelf life of the brochure is probably not that long, according to Colette. 'Developments are moving too fast for that. I suspect we will have to come up with updates at regular intervals. In fact, some domains are already building their own large language models such as ChatGPT for their own fields. This is already happening in legal. That tool will then be able to analyse laws and will know the difference between murder and manslaughter. If such a tool works well, does it still make sense to ask my students to learn the old-fashioned way from the law book? I suspect not. In short, the better the tools become, the more radical the changes that need to be made in education.'

Numerous questions

There must be countless questions about AI among Fontys employees. Wouldn't it be useful if Fontys itself set up an AI course for its own employees? Erdinç: 'We get that question more often. This item happens to be on the agenda of the next meeting of the AI workgroup. We are considering setting up a master class or something similar, based on the principle of 'train the trainer'. And maybe we will include the students in that as well.'

He summed up the working group's position: 'AI is here and it is not going away. We have to learn to live with it and work with it. But be aware that the generated texts are not always reliable. So gain experience with AI and research what it can and cannot do, and look carefully at your learning objectives and your way of assessment.'

Text: Martin van Rooij