How could I miss the latest HelF staff development session, as an avowed AI sceptic? Today Alice May and Shivani Wilson-Rochford from Birmingham City University talked about their approach to responding to the emergence of generative AI. As can be seen on the ‘roadmap’ above, this has included an AI working group, collaboration with staff and students on producing guidelines on use, sharing those via staff and student workshops, and collating resources on a SharePoint site. All things which mirror our approach at Sunderland.
Something they are doing which I liked was providing template text which academic staff can copy and paste into their assignment briefs on what kind of AI students are permitted to use, at four different levels from fully unrestricted, to fully prohibited. They are also working on an assessment redesign project which takes the risks of GAI into account, based on work from the University of Sydney which analysed all of the different types of assessment they have and put them into two lanes based on how secure they are to GAI plagiarism. It’s Table 2 on the page I’ve linked to, it’s a very good table. I like it a lot.
Briefly mentioned was the fact that Birmingham are one of the few institutions in the UK who have enabled Turnitin’s AI detection tool, and I would have liked to have learned more about this. From a student survey on GAI, the second screenshot above, concerns about the accuracy of AI detection was one of the big things they raised.
Alice and Shivani left us with plans for going forwards, which is to build a six-pillar framework on the different aspects of GAI’s impact on HE (third screenshot). Pillar 5 is ‘Ethical AI and Academic Integrity’. This one stood out as, once again, the ethical issues of the environmental impact and copyright were raised. Briefly. And then we moved on. It consistently bothers me, and I don’t have any brilliant answers, but I will reiterate the very basic one of simply choosing not to use these services unless they are solving a genuine problem.