Meeting of the Education Committee earlier this week (UK research university).
The student rep makes a passionate plea for us to deal with AI-based cheating in assessments. He explains how demotivating it is for the conscientious students to see their less conscientious colleagues boast about getting good marks for AI-written work.
In another section of the meeting, after the student rep has left, colleagues get all excited about setting up AI chatbots for student advice roles and designing AI-based marking tools.
The irony seems to have been lost on most of my colleagues.
The world we live in.
#HigherEducation #noAI #StopTheAICorruption #AcademicChatter
Noam writes likes this.
reshared this
Helen LH
in reply to the roamer • • •reshared this
the roamer reshared this.
the roamer
in reply to Helen LH • • •@Research_FTW
Well done! You motivate me to do the same.
The one silver lining is the persistent presence of these student concerns. Our students haven't given up on real learning --- not yet, not all of them.
David Chisnall (*Now with 50% more sarcasm!*)
in reply to Helen LH • • •@Research_FTW
You sound like an academic who understands the difference between formative and summative assessment!
I wish you were less unusual.
Helen LH
in reply to David Chisnall (*Now with 50% more sarcasm!*) • • •Helen LH
in reply to Helen LH • • •David Chisnall (*Now with 50% more sarcasm!*)
in reply to Helen LH • • •That's... not as surprising as I would hope.
Helen LH
in reply to David Chisnall (*Now with 50% more sarcasm!*) • • •David Chisnall (*Now with 50% more sarcasm!*)
in reply to Helen LH • • •@Research_FTW
I think that's somewhat understandable and the problem is that the rΓ΄les of teacher and researcher are conflated at universities (and the related problem is that the second is valued less than the first).
On paper, Cambridge has parallel tracks for research-only and research-plus-teaching positions. In practice, only the latter has tenure and most departments don't do the higher points on the track. This leaves you in a weird position where someone who wants to focus on research has to do teaching for advancement. At the same time, people who care about teaching have their promotion prospects curtailed by research.
And, looking at your bio, I realise you probably know a lot more about this than me, so I'm going to stop now.
Helen LH
in reply to David Chisnall (*Now with 50% more sarcasm!*) • • •David Chisnall (*Now with 50% more sarcasm!*)
in reply to Helen LH • • •@Research_FTW
Do you have any thoughts on the solution? I see value for having experts who are pushing the frontiers of their subject available to students, which implies some overlap between research-focused people and teaching, but at least 50%, maybe even 90% of what's taught in a modern degree course is fairly standard and doesn't get updated with recent research more than once every few years.
I wonder how much that has changed. Cambridge always likes to talk about Newton (who was a legendarily terrible teacher), and in his time Physics students would have learned a load of brand-new things.
And the same was probably true in the Maurice Wilkes era in computer science, when the department was building machines that were pushing the frontiers of the subject and students were directly contributing. But, today, you do a final-year project that is 25% of your final year, and that's the only thing where you get to do something completely novel.
That seems to be a fundamental change in the requirements of universities, funding models aside, and I'm not sure when it really happened.
Helen LH
in reply to David Chisnall (*Now with 50% more sarcasm!*) • • •David Chisnall (*Now with 50% more sarcasm!*)
in reply to Helen LH • • •@Research_FTW
Please let me know if you do! I would read it!
the roamer
in reply to David Chisnall (*Now with 50% more sarcasm!*) • • •@david_chisnall @Research_FTW
In most technical subjects, obviously we can't get students to contribute to the frontier, but we can teach foundational material in a way that opens doors to independent advanced studies.
Most importantly, we can help students learn how to "think as a researcher": explore possible claims, look for reasons that support or contradict the claim, try to relate the claim to other results. That can be done at all levels. Students in Year 1 often find this hard (why doesn't he just tell us?), but in Year 2 and Year 3, they increasingly appreciate the creative aspect of becoming a subject specialist. And, in my own field, in Year 3 I certainly can run a course based on reading classical research papers (as opposed to textbooks).
Are good researchers better or worse teachers? I have seen all combinations: the top researcher who's brilliant at exposition, or horrible at it. The caring & inspiring teacher without original thought, or original but unpublished.
Helen LH
in reply to the roamer • • •Noam writes
in reply to David Chisnall (*Now with 50% more sarcasm!*) • • •I'm surprised at teaching having priority over research at Cambridge. In my experience at UK universities, research tends to be more valued, with people who don't secure funding pushed into teaching roles, which are often thankless with unreasonable workloads. PG Cert and the like are mostly tick-box exercises.
the roamer
in reply to Helen LH • • •@Research_FTW @david_chisnall
Oh, everyone has taken these courses. You can get an AI feedback tool pass through all the standard pedagogical criteria.
These colleagues aren't stupid or ignorant, and almost everyone in such a committee will take their teaching very seriously. They will be good or excellent teachers. They just are caught up in the wave of genAI delusion.
the roamer
in reply to David Chisnall (*Now with 50% more sarcasm!*) • • •@david_chisnall @Research_FTW
I am not sure whether that distinction matters much here. All colleagues will understand the dustinction, and naturally most would be more concerned about cheating with summative assessment where the mark counts. To me, personally, using genAI in any type of feedback is damaging to the pedagogical process, regardless of the role of the mark.
Eg, letting students progress through automated feedback environments (formative) for most of the course and then assessing them formally (summative) by a human at the end to me is not the solution. The damage is done in the formative stages.
the roamer
Unknown parent • • •@alexpsmith
I don't think it's happening yet in terms of an actual freestanding tool. These were discussed as intentions, not current realities.
What does happen is that some colleagues let chatGPT help them write their feedback on an essay or a dissertation. They write in a style they have never used before. It stinks, both morally and pedagogically.
You mention the detailed comments you used to provide when marking. Ironically, AI tools are bound to be good at detail: they will pick up quotes from the student's text and make some deep-sounding but meaningless comment ("using more active verbs would add punch to your conclusion"), and they will do this at volume, filling in box after box of a marking rubric.
An experienced human marker may just write a handfull of comments, but they will assess the student's text holistically, as a more or less successful effort by the student to construct knowledge. One short paragraph of such human feedback is worth 5 pages of bot output.
the roamer
Unknown parent • • •@alexpsmith
Yes. I'm not ashamed to go all metaphysical on this. The human marker has a soul, and if the student's essay was written with a minimum of authenticity of engagement, however poorly expressed, then the human marker will recognise the student's soul in their feedback comments. Even if the essay is poorly written and the feedback comments are mundane, we have a true dialogue, a realisation of Martin Buber's "I and Thou". If the student lets the LLM do the writing, or the teacher lets the LLM do the marking, there is no dialogue, only role play.
Maintaining such dialogue is the rationale for university education, and thus it also supports the business case for a higher education institution. If we offer our students true dialogue and real learning, I am happy to defend the extreme fees we charge. Not so if we let bots do the talking.
the roamer
Unknown parent • • •@alexpsmith
The opportunities for such experiences are still there. Under threat, yes, but not squashed.
My daughters are 13yo so the choice isn't current yet, but if they were 18 this year and had a desire to study, I wouldn't hesitate to support this choice. I'd look carefully at where they apply, though.
πΈππππππ₯ ππ π¦πππ€ π
in reply to the roamer • • •