One year ago this November, ChatGPT-3 was launched and made widely available, and the world responded to the new chatbot with a mixture of excitement and concern. For the higher education sector, it was the risk of students using the new tool to help with take-home examinations that was seen as the immediate threat, although many also saw its potential. Here at Stockholm University, the President appointed a working group at an early stage. The group was tasked with quickly creating a guide to how teachers can address the issue of AI-powered chatbots, and with submitting proposals on how the university can continue to tackle the issue in the longer term.
The working group completed the first part of its assignment after just one month, and a guide is available here. The group reported on the second part of the assignment this autumn, and a few weeks ago the President decided on a number of areas where the university needs to conduct further development work.
One of these areas relates to making the university better able to follow reports of students suspected of cheating with the help of AI. So far this year, the Disciplinary Committee has ruled on eleven cases in which students were suspected of using a chatbot to formulate their answers. The reports and investigations have been thorough, and ten of these reports resulted in the students receiving a warning or a suspension. The number of unreported cases is probably very large, but these rulings nevertheless show that it is worthwhile for teachers to report when they suspect that students have not formulated their answers without unauthorised assistance.
Another area of development concerns the continuing professional development of teachers. Here, the Centre for the Advancement of University Teaching has been tasked with developing training and offering pedagogical support for the university’s teaching staff. Much has already been done in the form of seminars, workshops and consultations, both in general and for individual departments. The main theme of the activities has been how examinations can be designed in a way that reduces the risk of AI cheating. For example, take-home examinations – which will continue to be a permitted form of examination – should be supplemented with other types of examination to assess the students’ knowledge levels.
A third area of development has a more forward-looking perspective. AI-powered chatbots are here to stay and will be one of the many tools in the academic toolbox. This means that employers will assume that our graduates will have some competence in using chatbots responsibly to make their work more efficient. The President has therefore issued a broad mandate to review how we can develop our teaching and, where relevant, include elements that train students in the use of the tool. Of course, this also requires that we teachers have some understanding of the shortcomings and benefits of chatbots. We encourage you to test it out – if you have not tried it already!
This text is written by Clas Hättestrand, Vice President. It appears in the section ”Words from the University’s senior management team”, where the management take turns to write about topical issues. The section appears in News for staff.