Will students be allowed to use ChatGPT?
Eberhardt: ‘The policy doesn’t tell you what is and isn’t allowed. It provides lecturers and researchers with guidance and a framework. Learning objectives differ from one programme to another, so a one-size-fits-all policy wouldn’t work. For questions like these, we have a support system. Colleagues can contact the AI Office and the Centre for Expertise in Learning & Teaching (CELT). They have been working on various issues for quite some time. We were able to incorporate much of their preparatory work into this policy.’
Oolbekkink: ‘Safe and responsible use comes first. With ChatGPT, for example, there are genuine concerns about where data is stored. You shouldn’t just enter privacy-sensitive or commercially sensitive information into it. Your data may be used to train the system again, and students and colleagues need to be aware of that.’
Who decides what is and isn’t allowed?
Oolbekkink: ‘There isn’t one person who decides that. We focus on individual responsibility, awareness, and thinking carefully about what you’re doing. For example, when sharing privacy- or organisation-sensitive data. The frameworks for this are now in place. At the same time, as an organisation, we need to work on AI literacy among staff. We are developing a training course.’
Eberhardt: ‘Last year, the Executive Board already issued a statement on AI. The message was that we see it as our responsibility to contribute to the development and responsible use of AI. We want experimentation, but there are also risks and blind spots, for example in the areas of digital sovereignty, privacy, and sustainability.
Developments in AI also touch on a deeper layer of our identity as a university and that of our students, researchers, and lecturers. This policy is a first step. Our community needs to keep discussing the impact of AI on our university and on higher education in general.’
When does something constitute fraud?
Eberhardt: ‘That question cannot be answered in general terms. At UT, we have a quality assurance system, the Education and Examination Regulations (EER), and examination boards that consider individual cases. Colleagues with relevant expertise and developments surrounding AI are taken into account as we continue to develop this system.’
What will UT notice as a result of the new AI policy?
Eberhardt: ‘I think it’s important to approach this holistically. This policy now provides a framework for the use of AI in education and research. As a technical university with a social sciences faculty, we are fortunate to have a great deal of expertise in-house, in technical, ethical, and business administration fields. This policy was developed with input from faculties, institutes, and services and was also discussed with our partner universities. Our AI Office, the community organised by the Digital Society Institute, our colleagues at CELT: we have already been able to gather so much input, but we want to encourage everyone to contribute their ideas on how we can best develop and implement this further.’
How do you monitor all of this?
Eberhardt: ‘That is a difficult question, indeed. In education, we work through existing structures, at programme level and with examination boards, looking at learning objectives, teaching methods, and assessment. On the research side, our main focus in the coming period will be on awareness. The policy has also been incorporated into the R&I Charter that we introduced earlier this year. Responsible research is broader than AI, of course. The Netherlands Code of Conduct for Research Integrity was recently revised and is the guiding framework here.’
Oolbekkink: ‘We cannot monitor what everyone does at an individual level. It affects all kinds of aspects of our organisation, which is what makes it so complex. And we depend on people’s own basic knowledge and integrity, which brings us back to the importance of AI literacy.’
