Screenshot of a search query in Google and ChatGPT
A second session from the University of Kent on AI / ChatGPT, this time student-led. It was good to hear the student voice on these developments, and I found it reassuring that they are identifying the same issues and raising the same concerns as staff. More than one of the student presentations talked about how ChatGPT is already displacing Google and other search engines as the first place they are going to find answers. Like in the screenshot above, which shows the difference in results when searching for the temperature in Egypt from Google on the left, where you get a list of links to follow through, and ChatGPT on the right which provides a far more detailed answer in a written form. The problem is, as identified in one of the presentations but not the other, is that there is no way to verify the data which ChatGPT is presenting as truth. With the Google results you can evaluate the sources and verify against others; ChatGPT is a black box.
There was another good presentation from a student at Northumbria who has done some early research with students who have used ChatGPT to find out why and what they are using it for. The results being that they are mostly using it to check their own knowledge (problematic if you can’t trust ChatGPT to give you true answers), and to generate ideas. They are also using it outside of education to, for example, help write CVs and job applications. This makes me feel like we are always going to be on the defensive, reactive as one student said. While the education sector tries to grapple with the technology and debates banning it on one hand and thoughtful ethical use – while protecting academic integrity – on the other, other areas of society plough on regardless and heedless.
In another presentation a student talked about experimenting with ChatGPT to produce a response to one of their assignments and, echoing Margaret Bearman’s point from the teacher led session last month, they found that the result lacked critical analysis, and believe it would have been a clear fail.
I was pleased to note that ethical issues were being raised by students, though largely in the context of equitable access. With GPT 4 and priority access going behind a paywall, the students who can afford to use it will, and those who can’t will find they have another new and innovative way of being disadvantaged. How long before we see the first university or college purchasing a license for all of their staff and students?
Once again all of the presentations were recorded and are available as a YouTube playlist.Leave a Comment