Hi Navya, I'm fascinated by this (about to submit).

Is it common for people to live stream their defense of their thesis? Never come across this before!

I follow andy stapleton. Some of this vids are good, some not so much. You have to exercise your own judgement. Isn't that the case with anything you read or view?

But... 3rd level students are using AI in very novel ways. I know students are using AI to review old exam papers, assess probability of topics arising in exams and using AI to generate novel questions with complete answers, which they then reviewed, in a group (!), all as prep for exams.

If anyone is interested in this I can make the link available. The primary focus is on the Challenges-Problems-Solutions connections, which offers a map for anyone in Higher Education to think about how to respond to AI.

Thanks

Thanks for the link James, looks very interesting.
On equity of access, consider that some students will be able to pay for premium AI product lines whilst others from economically disadvantaged backgrounds will not. That will impact gradings unless it is account for in rubrics.

well perhaps it depends on your current level of skills in writing? For someone addressing writing in an academic style for the first time, or someone whose first language is not English or someone why has dyslexia, LLMs can offer real benefits. A skilled writer may find it less or not helpful.

I laughed at this... because we have a current and short lived orange weather warning, which may amount to 0-10cms snow in high ground. Everything in moderation!

LLMs have some utility, for example upload your own document and ask it to identify how to improve the text for a specific context, detailing the audience and purpose. Ask it to specify its rationale for the changes offered. This provides useful feedback, that you can adopt or discard...

AI is not an IT problem! AI is a pedagogical problem... (spelling never my strong point) and it is the faculty, including but not led by IT, who need to find solutions.

Institutional change even slower than social change, unless it harnesses the power of early adopters.

Samuel, only dabbled with OLS and rejected it in a current PhD design (cowards!) but I am intrigued by the argument of an inherent bias within an algorithm. Can this source of bias be applied to how GenAI operates and criticisms that it generates commoditized "beige" content? Thanks.