Browse Comments — Raw (as collected)

We have to find alternative ways to ensure the use of AI is limited to copiloting and not replacing the writing process. An approach, maybe, is to ask for regular review chapter by chapter and incorporating this in the assessment grading. Also a viva is a must and potentially has to carry a higher weighting where students unable to defend what they've written will give them out as potentially using AI irresponsibly. But it will take a lot of honest admission and "thinking out of box" and do away with some academic orthodoxy.

The interesting thing about "the floor has risen but the writing is boring" is that it maps quite neatly onto what happens when students don't have enough knowledge of academic writing to do anything with AI output except accept it. Students who understand argument structure, evidence use, academic register — they can push back on a bland paragraph. Students who don't will submit it. The problem was never really going to be stopped by detection. It's a question of whether students have the underlying knowledge to use these tools critically rather than wholesale.

A sensible option would be to make all dissertations 90% primary research and all use of AI generated comment referenced as not original work.

Coming from a completely different subject area, I agree. My perspective is that AI-proof assessments are urgently needed in order to de-incentivise using AI the rest of the time; it’s genuinely ruinous for education.

I think part of this discussion may be focusing too narrowly on assessment methods. In reality, we already have many alternative approaches available: oral exams, pen-and-paper assessments, practical activities, project-based work, and many others. We can certainly continue designing new assessment strategies adapted to the AI era. At the same time, teachers themselves can also benefit from AI to support assessment, feedback generation, and learning analytics. For this reason, I believe we indeed need to rethink teaching, learning, and assessment processes more deeply, but I also see many opportunities to achieve positive outcomes. My main concern, however, is not assessment itself. The real challenge is how we teach and support students so they can use AI effectively, critically, ethically and responsibly to achieve the best possible results in their work and learning processes.

Anna Yeadell I think we need to look at this as a gradual shift over years. The writing in 1960s phd theses that I have read, and the work done is massively different to today, less literature referenced, more descriptive language even in science, more philosophy and less clipped. Using AI and then raising the bar through new more rigorous expectations in grade descriptors might be the only way forward? We need to keep the philosophical thought. Training in that is key.

Helen Shaw The historical point is a really interesting one, though I wonder if the shift away from descriptive language was as much a deliberate methodological choice as a loss. Either way, you're right that something has been squeezed out. On grade descriptors, I think the instinct is sound but they tend to describe what "good" looks like without showing students how to get there. I totally agree that training is key. Philosophical thought has to be built somewhere in the curriculum, not just expected to arrive.

At one university where I used to work and mark dissertations, a few years ago, there was a case where a student submitted something obviously AI generated. It was obvious from the writing on the face of it and even more so speaking with the student. I refused to mark it and the case went to the academic dishonesty proceedings. And then got stuck. Could I *prove* that it was AI generated? The student swore up and down that it was their own work. I asked what kind of evidence would be acceptable. Even offered to enlist the services of a colleague who happens to be forensic linguist. At that time, no standard of proof was acceptable and the dissertation had to be marked, and eventually the student allowed to resubmit another AI generated version. It dragged on for ages. We definitely need alternative assessment methods. For now, I am thankful that I no longer have marking duties.

Ai can basically answer ass student assignments. Ticking time bomb

Time to reinstate examinations!

I have developed a set of prompts that assist students with the workflow of a research paper, giving inspiration in the areas of topics and research questions, getting them to generate content with AI and then use AI to help critique and improve on the ideas and analysis. Students so far seem to think that it’s the best of both worlds. They do produce better work but they feel they’re in control and are learning by going through the process. It’s a work in progress but I would be happy to share the prompts with anyone interested

Paolo Spada In the context, that makes sense.

John Reeks I have mixed feeling on the impact on education. I have used LLMs to teach myself a lot of new things. It is like having an unreliable teacher that injects random errors in what they teach. And in some tasks it can be prompted in a way it simplifies human cross-checking. At the same time when an undergrad student confesses they have not read anything since chatgpt came out, or when it is obvious that some PhD students are not progressing due to it, the danger is obvious. We need a new pedagogy. Banning and monitoring in various forms will be part of it, but also dedicated technology developed in the university (those that have the money are already doing it), maybe hardware based solutions (like the monitoring laptop I described), and teaching how to use it properly. I have more problems with the ethics of using commercial models. I want the university to have ethical locally installed AI we can use without exploiting workers, damaging climate, and contributing to a political economics system that is abhorrent. I think we can figure out the pedagogy, if the ethics is fixed. But Covid did a number on the tech autonomy of our universities that are now all captured by microsoft or google. So unclear the UK can drive it.

Manuel Caeiro Rodríguez I completely agree with your assessment of the pedagogical challenge. In fact my original post is not about coming up with solutions. Is about university machines letting us teachers try them. I am just an hobbist, we have colleagues in the pedagogy department whose life has been dedicated to pedagogy that can help way better than me. I am sure they have a bunch of cool approaches to try. The real challenge is the resistance to change and the lack of willingness/resources/risk tollerance to support change. We need changes in procedures that have been used for more than a decade and that are at the basis of existing business models. For example, I have the hunch that the remote assessment enabled a lot of commodification, that brought in lots of money in fields that traditionally had no money. It allowed to cut staff, and to rely less on expert staff, while expanding the number of students. The explosion of masters for foreign students that can barely understand English, was partially enabled by the take home assessment model. Those students would not survive a single oral, but without those students many of us will not have jobs.

Innocent Kasiyano (PhD) yes most of us are thinking in that direction. Written exams are a no brainer, and we are switching back. But it has impacts on curriculum, some very fun classes in which we teach how to do primary research will be harder to do (apart if those classes accept the use of AI, this is what I experimented in one of mine advanced classes this semester), and for the dissertation the exam is an oral and the UK sector has moved away from oral due to EDI concerns. Certain personalities over/under perform systematically in orals ceteris paribus. Maybe we can use new oral procedures, or special considerations procedures, to deal with those concerns, but some smart colleagues do not think so. We could also accept that the dissertation will be AI augmented and we teach how to do it properly. But we will need to restrict the domain on only the topics the department has strong expertise on. Lots of interesting trade-offs. Teaching and grading used to be non challenging activities, now figuring out this conundrum is one of the most fascinating puzzles out there.

I've at the pointy end of thesis season TBH the bland prose is getting to me. Looking forward to what happens next. And I hope students are able to retain the knowledge they are supposedly obtaining through writing a thesis.

They are going to start adding typos on purpose so that it looks human? So bland prose and typos next... 😀

Robert Studholme in my cloister/starship class I do exactly that after forcing them to read in class and argmapping in groups and restricting the domain to only a few papers they can use. And I agree that for some students that play along seem to work well and they quickly realise that good AI work is more time consuming than the old school approach. Because the helpful bit is adversarial mode and it basically asks the student to do an extra layer of work. But I have the nagging feeling that in my class they do it, in other classes they adapt it to become an anti-detection strategy. I feel a bit guilty at times I might be enabling significantly more advanced cheating strategies. It is also confusing the Academic Integrity process because a student can play dumb and say that the different confusing practices around AI, some people banning it, some people actively encouraging within guardrails like we experimented, has generated whatever cheating they are being accused off.

Paolo Spada I’m also using Class Companion to generate a revision trail showing how texts were constructed over multiple submissions. Its feedback is immediate and the majority of students seem to be very much in favour of writing with it. I’m currently developing a new version of an academic writing course that is tailored to the situation in my university, and integrates AI as a tool for students to interrogate and improve on. We hope to publish it as an OER soon

Helen Shaw in my class in which I have added it as a requirement 70% did it. I assume some asked AI to do it and then among them the smart ones cross-check well by reading the paper. But given the level of concept stretching I feel that 30% only did a mechanical check, they just searched the pdf for the existence of the word in the page, ending up with superficially correct references. The smartest one probably realised that is faster to do the check manually because doing the cross-checking well takes as much time as reading, or maybe hybrid, use AI, but then read. In the dissertation only the few students that took my class added pages everywhere, because the university asks for Harvard that does not have page references for indirect references.