Seminar 1: Finding the way â On generative AI use in doctoral research and education, Campus Valla, 5 May 2026.
This text summarises some of the main discussion points that came up during the interactive part of the seminar following the presentation. The summary is based on audience responses to a Mentimeter survey prepared by the guest speakers, Eva Ă
kesson and Rachel Forsyth, in which hypothetical scenarios involving genAI use were shared and audience opinions on the âcorrectâ course of action discussed.
Academic practice
Grounded in a case example involving fabricated references discovered in a successfully defended thesis, the discussion revolved around how to handle such a situation. When participants were presented with the options of âdo nothingâ, ârerun defenceâ, and âtake away doctorateâ, the vast majority (52) voted to rerun the defence. The point was made that incorrect references are not uncommon in thesis bibliographies, but for some participants the difference lay in the word âfabricationâ, which, they argued, raised further concerns about whether anything else might have been fabricated in the thesis. Similarly, others highlighted the importance of academic rigor and honesty and how such instances might undermine trust in academic work.
Additionally, some added that the options of doing nothing or revoking the doctorate both seemed too extreme, and as a result they had opted for the only middle ground which was to rerun the defence. Ultimately, a desire for more contextual and nuanced ways of responding to this type of situation was expressed.
Acceptable uses of genAI
Participants also discussed (non)acceptable uses of genAI - from students using it for thesis work to supervisors generating feedback to said thesis work. Across multiple scenarios there seemed to be no clear consensus on definite acceptable uses, with many Mentimeter responses clustering around âmaybeâ. However, participants seemed generally more tolerant towards using genAI for administrative tasks, for translation purposes, or as potential support for dyslexic or neurodivergent persons. Participants were less optimistic about using genAI for direct academic tasks or for supervisory specific tasks such as generating feedback for drafts of thesis chapters or preparing opponent material for defenses.
The conversation also reflected some epistemic and disciplinary specificities. While the Mentimeter responses were anonymous, the discussions that followed indicated that what is deemed as (non)acceptable use of genAI, e.g. when drafting a literature review or summarising your own work, varies significantly based on one's disciplinary background.
Data ethics and privacy issues
Concerns about data ethics and privacy issues came up consistently throughout the conversations, ranging from scenarios of supervisors uploading student drafts to generate feedback to persons using it for medical and mental health support. Despite data fed into LiU systems is not to be trained on by genAI, some distrust was expressed in relation to whether this would in fact never happen.
The promise of more time
Another topic of debate surrounded âthe promise of more timeâ. Optimistic perspectives argued that these tools could be a clever solution to time-consuming and tedious administrative tasks and thus help free up time for what academia is really about - learning, thinking, engaging with your peers. This was however met with a degree of scepticism by some participants who expressed concerns over whether this was in fact realistic, especially within academia. Where would this new-found time actually go? Participants discussed the notion of having to move faster, meet higher expectations, and the risk of burnout.
Broader societal impacts
The broader impacts of genAI were also a topic of debate. Concerns about work displacement were brought up with participants discussing whether this was a natural development of modern society - arguing that new types of jobs would be created accordingly - or whether it begged for more critical attention to the impacts of these technologies and how they play a role in our lives.
Additionally, sustainability and the environmental impacts of genAI was also raised as a genuine concern. Here, the conversation revolved around the severity of the environmental impacts (compared to other IoT technologies), questions of when it was justifiable to use genAI (e.g. for scientific purposes vs. personal use), and others expressed a need for increased scrutiny towards these technologies and the infrastructures they sustain (See also Seminar 3 which will continue the conversation on sustainability in November 2026).
