University of Twente Recognizes AI’s Potential While Acknowledging Risks

University of Twente Recognizes AI’s Potential While Acknowledging Risks

(IN BRIEF) The University of Twente (UT) acknowledges the opportunities presented by Artificial Intelligence (AI) tools like ChatGPT while emphasizing the need to be aware of associated risks. UT believes that AI can be a valuable tool in education and research but underscores the importance of training teachers to handle and use AI effectively. UT emphasizes the human factor and critical thinking, as AI is not flawless and may contribute to educational inequality among students. UT highlights the need for AI literacy among teachers and students and emphasizes that human involvement is crucial for responsible and ethical AI implementation. The university recognizes the ethical concerns, including affordability and equal opportunities, associated with AI tools and advocates for investments in large language models for common use. Additionally, UT highlights the potential bias inherent in AI systems, as they rely on human-provided information, which can distort reality.

(PRESS RELEASE) ENSCHEDE, 19-Jul-2023 — /EuropaWire/ — Tools like ChatGPT have become indispensable in education and research. Therefore, the University of Twente (UT) embraces developments in Artificial Intelligence and embraces the opportunities that this technology offers. Being aware of the risks is also crucial, as the human factor remains paramount: studying and conducting research are human endeavours.

OPPORTUNITIES AND RISKS

AI such as the ChatGPT tool provides us with opportunities. According to professor Kim Schildkamp and educational consultant Robin van Emmerloot at UT, there is no need to panic about its rapid development: it may actually be a valuable tool in education and research. It is important to train teachers adequately, so that they and their students learn how to handle and use AI. The human factor remains indispensable, as does a critical mindset, because AI is not flawless and there is a risk of educational inequality among students.

AI SKILLS

The speedy rise of ChatGPT quickly sparked a discussion about AI development. Kim Schildkamp, professor in the Department of Teacher Development, was not surprised by the widespread adoption of the program. “We have been working on this topic at UT for a while and we saw a development like this coming. I understand people’s concerns about AI, but I believe that the main question is: how do we use it to enhance education?” Similarly, Robin van Emmerloot, an educational consultant at the Centre of Expertise in Learning and Teaching, considers the upheaval exaggerated. “This form of intelligence will remain artificial and these tools are by no means flawless, so we do not need to worry that students will not have to put in any effort to obtain a diploma. Instead, there are challenges for students and teachers in exploring how these tools can help them: which AI skills does someone need in an AI-driven society?”

TOOL IN EDUCATION AND RESEARCH

Using AI requires a critical examination of the information that is gathered, but it can also be a useful tool in education and research, according to both the professor and the educational advisor. Schildkamp states: “ChatGPT can be utilised effectively as a tool to achieve certain learning objectives. If a learning goal is focused on something other than thinking skills such as correct spelling and grammar, AI can be helpful. In research, AI can assist one in tasks such as writing a research proposal. Transparency is essential, both in education and research, so using AI tools should always be made public. Additionally, it is crucial to check the sources provided by the tool thoroughly, as sometimes the program simply invents a reference.”

TRAINING TEACHERS IN AI

Educating students starts with the knowledge of teachers. “Teachers can still learn a lot in terms of AI literacy. They need to understand how to make use of it to support their teaching and to help students learn, as well as the meaning of AI literacy for students’ future professional fields. Therefore, educating teachers and students in this area is a priority,” Schildkamp emphasises. Van Emmerloot adds: “As a teacher, you can also use AI as a kind of teaching assistant, for data processing, creating scoring rubrics, and setting up your e-learning environment, for example. In such cases, AI can act as a co-pilot, while you remain in control.” UT has developed a protocol for the use of AI that highlights the importance of human involvement. The Human Factor

To use the technology effectively, humans are indispensable, as is emphasised by the University of Twente’s Executive Board: “We must embrace AI technology cautiously and strengthen the human factor in education to handle the technology responsibly and ethically.” Schildkamp adds to this statement: “In order to be able to learn well, a sense of community is essential. This has become evident from research conducted by one of my PhD candidates, Linlin Pei. Students need to feel that they are part of a group and they need to have a relationship with teachers and fellow students, because these conditions contribute to their ability to learn. Obtaining a degree for teacher training or engineering is unlikely to happen solely with the use of AI. You still need teachers and peers, you need this sense of community.”

THE ETHICAL ASPECT

In addition to the opportunities that AI offers, the ethical aspect concerns Schildkamp. “ChatGPT has a clever marketing strategy with a free initial version, but the paid version is already capable of much more. Not everyone can afford to pay for that,” she explains. Van Emmerloot agrees, drawing a parallel to the situation during the COVID-19 crisis: “Students with good internet connections and computers were able to deal with remote education better than those who did not have those facilities at home. Although it may not be an issue yet, equal opportunities are certainly threatened by the price tag that is attached to premium versions of AI tools. For educational institutions, one way to address this issue is to invest in large language models for common use.” Schildkamp adds: “Furthermore, we have not even discussed the bias that is inherent in these types of AI systems, as they are based on the information that we humans provide them with. This means that the information that is offered can present a distorted view of reality.”

Media Contact:

J.C. VREEMAN (JOCHEM)
+31 6 5432 6671
j.c.vreeman@utwente.nl

SOURCE: University of Twente