Scotland Staff by Sue Attewell

2025-06-25 12:37:14 UTC

At Edinburgh, we have a home grown interface to some of the learning models that we're encouraging staff and students to use - ELM - (Edinburgh (access to) Language Models) | Computing | Information Services

We are advised to use CoPilot and also have Grammarly -Queen Margaret University

2025-06-25 12:44:51 UTC

AI or GenAI

2025-06-25 12:45:05 UTC

Looking at both AI and GenAI, at Abertay, there's also Padlet's GenAI and Read and Write's AI voices which I don't think have been mentioned yet.

2025-06-25 12:46:31 UTC

At St Andrews, Copilot has been made available, but I don't believe we have a curated list of other "approved" AI softwares, as some of the others have indicated.

Ethics

2025-06-25 12:47:06 UTC

So many issues, some raised by Dan McQuillan here

Hype

2025-06-25 12:48:24 UTC

People falling for the 'awe' of supposed AI capabilities and not understanding how much of it is hype - all the while pushing it forward in their institutions

Strathclyde

2025-06-25 12:49:17 UTC

We have had a very limited pilot of CoPilot, and limited projects that have been able to use Open AI APIs either via OpenAI or MS Azure version. Also run some very constrained local LLMs as demonstrators.

But no "wide" availability really, primarily linked to the costs

Trust

2025-06-25 12:49:51 UTC

Risk of break down in trust between students and staff. Especially when it comes to assessment. Whether or not suspicions are justified ... it's the impact on trust in the educational relationship that's the risk.

UHI & UHI Inverness

fionamcconnellic — 2025-06-25 12:49:56 UTC

These will all have exclusions/T&C and age limits for staff, students and apprentices to agree to - We are a MS Office place so main one is MS Copilot in Bing, AI Tutor Pro, Claude.AI by Anthropic, Elicit, LeonardoAI, OpenAI ChatGPT, Perplexity, Phind, Satlas, SciSpace, Semantic Scholar.

Impact on skills development

2025-06-25 12:50:36 UTC

Keep hearing chat about how generative AI can adequately undertake literature reviews, synthesise and summarise - lack of engagement with research evidence about how they can't (and functionally will never be able to)

AIAS

2025-06-25 12:50:54 UTC

maybe not the intended question but from 25/26 we will be using the AI Assessment Scale for all taught programmes

2025-06-25 12:51:01 UTC

Free versus premium and the gulf between them.

2025-06-25 12:51:14 UTC

I concur with the issues raised by colleagues - big concerns re: ethics, sustainability, the accuracy (or lack thereof) of outputs, potential for reinforcing mis/disinformation. But at the same time, there is a need to teach students to use it appropriately.

GCU

2025-06-25 12:51:19 UTC

We recently had an AI Framework which establishes basic Co-Pilot in Office 365 as the tool that can be used (the exact boundaries are yet to be established), and any public AI tools are can only have publicly available information added to them. We are currently conducting a data protection review of the AI Design Assistant in Blackboard Ultra and we will look to use this when we fully adopt Blackboard Ultra.

The AI Tools at GSA:

2025-06-25 12:51:43 UTC

Copilot (not the full license), Adobe CC/Firefly, Otter.ai, Grammarly

Some of our programmes use specific tools: Midjourney; Runway

Intellectual laziness!

2025-06-25 12:52:00 UTC

Increasing amount of research around how people using AI quickly become intellectually lazy, less motivated or able to think critically etc

2025-06-25 12:52:31 UTC

The panic around detection and the potential for student data being added to external tools with no data protection review.

Fear about resistance

2025-06-25 12:52:49 UTC

People who want to resist the use of AI (because it's not inevitable) don't feel like they can because of pressures from higher-ups

fionamcconnellic — 2025-06-25 12:52:52 UTC

Students new to using AI, thinking it a fountain of knowledge and not lecturers/tutor etc.. - taking what it says for granted, not fact checking, understanding the biases exist and the differences between these in the many different AI tools that are out there.

Not Really

2025-06-25 12:53:06 UTC

Largely self-taught

2025-06-25 12:54:37 UTC

Not initally, but I think many of us are learning as we go and asking better questions about it.

2025-06-25 12:54:51 UTC

Could certainly do with more development in this area - but finding the time is challenging! We're introducing workshops for staff for the first time this coming academic year.

Critical

2025-06-25 12:54:56 UTC

I want to see higher education that has standards - supporting students to be critical members of society, upholding ethical standards, choosing not to use tools owned by some deeply problematic characters and that are burning the planet

2025-06-25 12:55:23 UTC

New AI features appearing in existing platforms. We want to have some control about whether these get turned on or not based on whether they seem useful and/or have the relevant contractual changes made. We've developed a process to manage this: Scoping AI in EdTech at Edinburgh (SADIE) | Learning Technology | Information Services

Yes/No

2025-06-25 12:55:32 UTC

Hard to keep up and I have not started my Adobe Express training yet so maybe that will improve my work in ways I don't know yet.

Yes

2025-06-25 12:55:43 UTC

I know plenty about it and how to use it and choose not to use it because it's not good enough and is deeply unethical

Equity

2025-06-25 12:55:44 UTC

Access to tools
Uneven approaches to assessment module to module (even within same programme)
Uneven approaches to academic integrity between faculties

fionamcconnellic — 2025-06-25 12:56:36 UTC

I started my AI journey beginning of 2023, I am in AI tools every day, I am on socials that follow AI folks, experts ( so called), and I still feel overwhelmed by it all. Call me sinical, I suspect that is a deliberate ploy by people/companies as its a money maker, by those folks 'selling' you support, how to create AI Agents etc, to companies getting you to take the upgrade with the latest AI [insert whizzy LLM}..

Assumptions around workplace use

2025-06-25 12:58:04 UTC

An argument for providing AI and training at university I keep seeing is that workplaces will be expecting it and using it - evidence suggests this isn't the case, it's not being effectively used in various industries, and it's frustrating that this is used as supposed justification

2025-06-25 12:58:12 UTC

The way that higher education is dancing to the tune of these incredibly powerful, rich and frankly immoral companies. The discourse is totally normalising hugely harmful technologies, and we as a sector are amplifying that by saying it's the future of the workplace and we should be preparing students. It's completely uncritical and antithetical to the common good of higher education.

64 million dollar question

2025-06-25 12:58:13 UTC

Need real time with academic colleagues to explore opportunities for assessment re-design

2025-06-25 12:59:08 UTC

Values-led, skills-led, flexible for UG & PGT...

Academic Integrity, Cognitive offloading, Employability vs AI replacing employment opportunities, Ethics (sustainability & training GenAI Assistants), lack of transparency/black box.

2025-06-25 13:00:16 UTC

2025-06-25 13:00:47 UTC

It's hard to benchmark what those skills are, so unsure how to know if I have the necessary skills. It's quite a challenge.

No

2025-06-25 13:01:11 UTC

I don't have the specialist knowledge to truly upskill my academic staff on how they might really use genAI in a meaningful way in the curriculum, and I don't have time personally or resource in my team to acquire this knowledge and stay current with it in this fast moving environment. So I am frustrated.

Disability-washing

2025-06-25 13:01:50 UTC

Really concerned that arguments about how great AI is for disabled people (e.g. ADHD, dyslexia, physical disabilities) is just disability-washing a harmful tool that's actually potentially very bad for disabled people

Bad advice

2025-06-25 13:03:25 UTC

Keep seeing examples of students' work where they've been recommended to use Grammarly - it's giving really bad advice! I tried some of my own work on it and it couldn't make sense of complex sentences, couldn't agree verbs, suggested inappropriate synonyms etc

Yes-ish

2025-06-25 13:03:42 UTC

But as an IT Developer who has "grown" an educational aspect, the training / develoment for Education sometimes "feels" more hyperbolic, than useful vs. the "technical" or pratical type training you see for developers to build AI tools.

Supportive of intellectual property and ethical use of information

2025-06-25 13:04:36 UTC

Not supportive of intellectual theft from generative AI tools, which are (all? mostly?) built on stolen content

Is ChatGPT taking all jobs?

fionamcconnellic — 2025-06-25 13:05:31 UTC

I am concerned about the hype in the press, that mention jobs going, and linking in that its generative AI - whilst I do not know anyone in these companies, I do not believe for a second its a free/basics AI versions, of tools ChatGPT, Claude etc that is taking all of these jobs, likely its automation using AI Agents ( at a high cost that these companies are choosing to buy) with more pro AI software, that has been programmes by a human - people using ChatGPT, CoPilot etc isn't going to cause 1000s of job to go.

With a spine

2025-06-25 13:05:34 UTC

I think HE institutions should take the stance that generative AI isn't ethical for many reasons and as a result they don't support its use

Support to resist

2025-06-25 13:09:44 UTC

Creating the conditions for people to feel like they can say no to using it without being accused of being a luddite.

Squaring the circle of values

2025-06-25 13:11:02 UTC

Institutions concerned with sustainability, social justice etc. and how they can walk that walk while encouraging use of socially and environmentally damaging technology

What are the skills?

2025-06-25 13:11:19 UTC

It depends on what these skills are. I'd say for everday use it may be straighforward, but what academic skills looks like for Gen AI

In society

2025-06-25 13:13:20 UTC

Some forms of AI are demonstrably making improvements to healthcare (identifying cancer early etc) but this is fundamentally different to what's available in HE, when what we're really talking about is LLMs. LLMs aren't improving anything

Digital asbestos

2025-06-25 13:15:10 UTC

How we're going to fix the internet when generative AI has fed so much slop onto the wider web that it means we can't trust the content we encounter

2025-06-25 13:16:14 UTC

One of the primary skill required is a critical approach to using the tools; deciding if and when they even need to be used, and then how best to use them.

For talking out loud/ideas generation

fionamcconnellic — 2025-06-25 13:16:37 UTC

love it for sorting out my brain dumps of nonsense and limited use of commas and full stops. Used it to map against standards, for training as a starting point to allow me to step back and see what I had done, compare and review. This quicken up the process by hours, freeing my time for other tasks. I never use it summarise and even personally looking for ideas for a recent trip, it made up pubs/ restaurants to visit, with some 20Km from the local point I asked it to map it to.

Training the AI-pilled

2025-06-25 13:18:36 UTC

I often encounter colleagues who think AI is working for them and I'm very concerned that they genuinely don't realise it confabulates, can't fact check, can't adequately synthesise information, has bias baked in - some with the Eliza effect in full force. It's hard to know how to challenge some of the misunderstandings about the nature of AI, particularly when there is a power differential

2025-06-25 13:21:54 UTC

The ability to evaluate evidence and sources is also key - users need to be able to critique the outputs of AI to decide if they seem plausible on the face of it.

More 'global' level conversations about where we think this is going. It seems that the tech companies are changing our world in a way that the majority of people have no say in.

2025-06-25 13:22:08 UTC

subject-specific authentic, really?

2025-06-25 13:23:16 UTC

I guess "authentic" assessment now should be re-defined in light of AI and graduate skills expectations. This is changing heavily across the disciplines.

Assessments and conversations

fionamcconnellic — 2025-06-25 13:23:27 UTC

Without people able to have honest conversations with students, asking if they wrote X assessments, and students not lying about AI doing the assessments, or even family members, essay mill etc.. Then I suspect, sadly it will either move to closed book exams or would be open book assessments for those that can. The above is from experience, where now I am seeing more students having family members doing their work and they (both students and family), I also suspect they are also the ones encouraging them to use AI tools in their assessments and think that is ok to do - a broader conversation in society is perhaps needed that its ok to fail, not get the outcome you want, including some responsible needs to be on the students - learning is hard, keeping to deadlines is hard along with work/family commitments. Its up to them to do the work with support from lecturers/PAts etc.

2025-06-25 13:24:20 UTC

Speaking back to society about GenAI, leading the debates, informing, educating widely - not just our students - about the ethical implications of these tools. Taking back ownership of our expertise and as a place where knowledge and learning is valued and protected.

A portfolio of secure and open assessments. Authentic assessments and 'process' type assessments.

2025-06-25 13:26:36 UTC

Assessing the process not the output

2025-06-25 13:27:58 UTC

Assessments should focus on assessing the process not the output

What are we assessing?

2025-06-25 13:52:57 UTC

1) Having conversations with colleagues about whether to assess use of AI, particularly in a disciplinary context. And the importance of any such assessment linking to their learning outcomes. But we need to think about whether this is appropriate? (i.e. some disciplines it's important to be able to use AI, and some it's not or it's being included because it's the current hot topic)

2) And then ongoing conversations about authentic assessment (although this started well before GenAI became widespread). Thinking about what we are assessing and is it really authentic?

Information and digital literacy training

2025-06-26 10:12:26 UTC

Rolling out of something like this course, which can help students (and HE staff) understand the principles of AI

Impact on learning/skills/competence

2025-06-26 10:23:31 UTC

There's an increasing amount of research coming out about the impact of use of AI on people's ability to learn, meaningfully engage, apply critical thinking etc. This has implications for students but also in healthcare sector and other sectors.

One study showed that the diagnostic ability of radiologists plummets when they are working in concert with an inaccurate AI.

Misinformation and its impact on academic research, student assessments, wider social impact

2025-06-26 10:25:34 UTC

A number of AI-generated and error-ridden books about mushroom identification have been found on Amazon. In a viral Reddit thread, one man claimed that one family ended up in the hospital with severe poisoning after buying and using one of these books. - apply that to an academic context without adequate screening, curation, evaluation etc. - scary