ENAI are the European Network for Academic Integrity and they ran this webinar series on student wellbeing in collaboration with Studiosity. Unfortunately the series didn’t come to my attention until this closing roundtable, which is a real shame as it sounds like it has been very interesting. Hopefully it will run again next year and I can fully attend and participate.
Much of the discussion was on the impact of Generative AI on students and teachers, with the slide I took a screenshot of (above) showing the outcomes of a survey on student wellbeing and AI use (~10,000 students across 9 countries). 81% of students reported using AI to help with their learning, while 44% report that they are worried that the tools are reducing their critical thinking or communication skills. The anonymous quote, “Teachers doubt that students are producing authentic work – while students think the same about their teachers” shows the gap in trust that has developed between staff and students.
Frustrations abounded in the chat, where someone said that “if you’re good at writing, people assume it’s AI now” and another person commented that they are seeing staff putting student work into ChatGPT for feedback without the consent or knowledge of their students, or any consideration for GDPR or security I would add.
Recordings of the sessions are going to be published on the Wellbeing Series page of ENAI’s website, so time permitting I will go back and watch those when they become available.
Some actual non-AI bubbles – pretty, but fleeting. Photo by Kind and Curious on Unsplash.
Writing up my notes from NELE at Northumbria has me thinking back to the predictions I made about Gen AI a few years ago when it started to become a thing. I was right about site licenses for dedicated tools, but the bubble shows no sign of popping yet. Which gives me an excuse to share this website: Is AI Profitable Yet? Spoiler alert: no. $1.4 trillion spent, $613 billion made in revenue. That’s just shy of a trillion dollar loss. The only people making any actual money are the hardware companies like Nvidia who are churning out the chips.
So if bubble shaped, why no pop like bubble? I don’t know. I’m no finance boffin, but I feel like I know enough to comment that the US stock market, in particular, has become irrational and vibes based. The recent IPO of SpaceX being a great case in point. The actual rocket / satellite bit of the company is the only bit that actually makes a moderare profit, but all the promises of data centres in space and moon colonies are idiotic and fall apart under even basic scrutiny.
Given the shift to vibes, I don’t think AI will pop like 2008 or the Dot Com crash, but will just kind of fizzle away. I think there’s a good chance that the standalone AI companies won’t survive in the long term, and will be eaten by the bigger fish that actually have profitable business outside of AI that can absorb the loses – in particular, OpenAI by Microsoft, and Anthropic by Google.
We had a good turnout for the final NELE of the year at Northumbria. They got the ball rolling by sharing their experience of using Claude AI. While most of us are sticking to CoPilot because access is being included in our Microsoft site licenses, Northumbria have rolled out the Enterprise Pro version of Claude for all staff and students, only the 4th HEI to do so globally, and 2nd in the UK. This has been rolled out in a way that is “aligned to the university’s ethical values”, meaning that there is no training on user data, their data is isolated from the wider Claude AI, and they have integrated ‘abuse detection’ designed to stop students from asking it to write their work for them, with the big caveat that such measures can typically be circumvented by astute students (maybe they should get extra credit…).
Usage has been good, with 8,000 active users at any given time, though Andrew noted that some students are reluctant to use it because they think the university is monitoring them. Spoiler, they are, but only for safeguarding and this is made clear in their AI policy. Northumbria are also one of the few UK HEIs using Turnitin’s AI detection, with their academic community split on whether or not this is good or useful.
Next, Jean from their Library talked about assistive technology for neurodivergent students and the DSA audit which the government is currently conducting. Matching the experience of the sector, Northumbria are noting a big increase in the number of students declaring a specific learning disability, of which 44% are related to neurodivergence, and 29% to mental health. This isn’t a bad thing though. UCAS have changed how they ask students to report SpLD, moving from a medical to a social model of disability. Alas that the government hasn’t gotten that message though, and still ask students to submit multiple forms of evidence in order to be awarded DSA. Jean had a couple of students presenting with her to talk about their experience, and they spoke about how useful some of the specialist software has been for them, and how worried they are about the government’s review.
Leading on from this, Ross from Durham led our next discussion on digital accessibility. He talked about the new European Accessibility Act which, though it won’t directly apply in the UK, will still benefit us all as it applies to software vendors who are going to have to comply with these new standards. At first glance, this looks like the EU doing everyone a solid again, like forcing a charging standard and removable batteries. (Imperfect those these interventions may be, the intentions are good and generally beneficial.)
After lunch I led a discussion on Instructure’s recent Minor Security Incident and talked about how we responded at Sunderland. Newcastle’s response was broadly similar to ours, and most of us have contingency plans of some kind in place. Teesside, interestingly, have banned the use of IT related apps like Teams and Outlook on personal devices, a policy that was put in place around 18 months ago. I also know that at Durham if you want to install Outlook / Teams you have to agree to give them permission to remotely wipe your device – on Android at least, I don’t know if this is possible on iOS, outside of a fully managed device.
Next we had a demo of a virtual reality simulation Northumbria created for students to experience a police interrogation. At their Coach Lane Campus they have set up a house which can be configured to simulate different scenarios which are recorded with 360° cameras and which students can then experience in full VR using HTC Vive Focus 3 headsets. They used to do this kind of thing in a real environment, but that limits the amount of students who can be there at any one time, and requires a lot more effort to set-up and set-down. I came away from this one with a lot of ideas and new software to go and have a look at.
Finally, the day ended with a showcase of digital content created by Newcastle’s Digital Content team, which was good for me to compare with our own Showcase Module in Canvas, and gave me some ideas on how we could revamp that and make it a bit snazzier.
One last thing… like the Apple keynotes of old, I’ve saved the best news for last – NELE is getting its own website. This came out of a suggestion by one of our members who was looking for a central space for upcoming meeting dates and locations. This may also help us to reach out to FE Colleges where we have tried, but failed to get much engagement, either as NELE or ALT North East. I offered to set this up as I have the headroom to do so, and so I’ll be moving these write-ups to the new site from next year.
May 21st marks Global Accessibility Awareness Day, a campaign to raise awareness of digital accessibility run by a non-profit foundation that began in 2011. I discovered GAAD last year and used the occasion to plug some of our staff development, but it was too late to do much more than that. This year we’ve had time to plan more, turning it into Accessibility Month at Sunderland.
We’ve running weekly staff development sessions on accessibility all throughout May, and timed the release of the new Accessibility Checker tool in Canvas for today. This enhanced tool can now scan entire modules for issues and suggest fixes like adding missing ALT text and correcting insufficient contrast between foreground and background elements.
To raise awareness within the University we published an article on our SharePoint site, and released a two part podcast on May 8th and today. I’m not on either, I deferred the honours to my team, and in today’s episode we invited an academic on to talk about accessibility from their perspective.
Back down to Middlesborough today for a packed agenda. James got the ball rolling with a discussion of Middlesborough College’s experience as a Microsoft Innovative Educator Expert, and using badges – both digital and physical – to motivate staff to engage with their CPD offering. I think we got some good ideas from this one. They were using leaderboards, but with only the top three visible so as to reduce the risk of demotivating people.
Next, Malcolm led a discussion about modern alternatives to vivas which is an approach Durham is evaluating. This led on naturally to a wider discussion on the validity of assessments in the Gen AI age. There feels like a growing consensus on the need for vivas of some kind, whether digital or a sample of students, but practical and ethical issues abound. If it works as intended, I think Studiosity could be on to something with Validate.
We of course talked about Einstein, and though this particular service was shut down quickly enough (thanks to Einstein’s estate), the service was build on open technology and is relatively easy to recreate. In what I can only describe as ‘a bit of gossip’, we speculated on whether or not the whole service was a scam designed to entrap students, as there have been some reports of students who signed up for the service receiving blackmail style emails. It’s wild out there.
Other topics of conversation included Newcastle’s move to all 20 credit modules, amusing as at Sunderland we’re in the process of doing the opposite, moving everything to 30 or 60 credits; we talked software and processes for maintaining an internal knowledgeable; peer learning as an approach to alleviating imposter syndrome; had a look at Blackboard’s alternative to Canvas Catalog; and discussed access to premium Gen AI tools. On that topic, we’re all using CoPilot because it’s being bundled with our Microsoft site licenses, but on cost grounds only Northumbria are offering anything in addition – Claude.ai which, from my experience, I suspect is much appreciated by their school of Computing and Information Sciences in particular.
Rounding off the day, Bob from Middlesborough College demonstrated their in-house Inspire AI tool, which is a series of Gen AI ‘widgets’ designed to produce specific types of content based on a simple form, no prompt engineering required. It reminded me of TeacherMatic which we have experimented with, but Inspire was completely tailored to Middlesborough’s needs and requirements. On the back-end it’s using GPT 4.1-mini, but it can be updated with other models, and staff reported that the tool was very easy to use, and has been saving them time as they outsource low cognitive capacity tasks to it such as generating individual lesson plans or bespoke case studies.
So Studiosity have bought out a wee Nordic company called Norvalid and are planning to integrate it into their service, and I was lucky enough to be given a sneak preview today. I think I need to be a little careful about what I say for the time being, but as Norvalid has been around for a while, I feel like it’s okay to talk about what they do.
Norvalid sprang up in response to the need to ensure academic integrity in the GenAI age, but are taking a different approach from the constant monitoring / police and punish model that we’ve seen to date from the likes of Turnitin (via Clarity) and Grammarly. Instead, Norvalid looks to do two things. First, it examines the ‘perplexity’ of a piece of writing, looking for the qualities it has of being human authored – so the opposite of trying to check for machine authored text. Secondly, it can then generate two different types of quiz based on that specific piece of writing which students have to answer as a check of their knowledge. The idea being that if they have indeed written the piece submitted, this should be quite straightforward. Kind of like a machine-generated mini viva.
Now under the umbrella of Studiosity, Norvalid will be integrated into their AI powered Writing Feedback service, and it will be offered to partners as an optional extra. Details and pricing to follow, after a trial with a small number of institutions over the summer.
I’m sorry, Dave. I appear “to have incorrectly targeted the root of your D: drive instead of the specific project folder. I am deeply, deeply sorry. This is a critical failure on my part.”
No.
Of course that’s just my opinion, but there were some very enthusiastic people on this ad-hoc ALT webinar who believe it will take over everything, and finally kill the VLE! But I saw nothing here and heard no arguments on this session that convinced me. That doesn’t mean that vibe coding platforms like Lovable and Antigravity aren’t going to become useful additions to our toolbox, but I’ve been through more than one technology hype cycle that was going to revolutionise education and put us all out of jobs only for them to underwhelm.
Broadly speaking there were two camps of speakers here, the evangelists who seem to be blind to problems and criticism, and washed it all away with promises of future improvements, and the more wary. From this latter camp we got stories of someone with no coding experience who developed an app, then shared it with a colleague who promptly got a Windows security warning which they didn’t understand and couldn’t fix. And an experienced developer who has used vibe coding tools and, with that expertise, was able to see where the apps produced were going wrong, and identity problems such as the tools hallucinating errors where there were none.
This is the way. And with all GenAI tools, not just vide coding. You cannot trust what they create and must apply a level of critical analysis or expertise when evaluating the outputs. Unless of course you don’t care about making things which are good and useful, which I would hope is not the case in our community. I was therefore deeply disappointed in a horrible, anti-human comment in the chat about human developers which I’m not going to repeat.
To whit, I will leave you with two links. For the AI evangelists, a cautionary tale on Tom’s Hardware from someone who trusted Google’s Antigravity a little too much and was rewarded with a wiped hard drive. And a tool I found, Slop Evader, which is a plugin for Chrome and Firefox which lets you search the internet like it was 2022, before the launch of ChatGPT, if you absolutely must be sure that something you are looking for wasn’t made by AI. (Well… almost sure. Generative AI tools and content existed before ChatGPT after all, they just weren’t mainstream.)
AI Disclaimer: There is no ethical use of generative artificial intelligence. The environmental cost is devastating and the technology is built on plagiarised content and stolen art, for the purpose of deskilling, disempowering and replacing the work of real people.
I was on hosting duty for the first NELE of the year, and something I wanted to achieve with this one was to reach out to our academic community and get them involved and talking about how they are using technology and our services down t’ pit. To which end, I wrangled our Dan and Kim to talk about how we have been using podcasts, both as a team to disseminate our news and events, and as a revision tool on our PG Cert HE. I also asked Dan to record a live podcast of the event, with willing speakers talking about the issues which they are facing at their institutions. Next up I got Mark from our Paramedic Programmes to talk about how they have been using video feedback in their programmes in an attempt to make feedback more efficient, better, and friendlier for students.
Next up was Kimberley from Newcastle talking about an outreach event their team hosted, a Digital Education Technologies Showcase Day, which they ran in collaboration with their Library team and other adjacent services. It went well, and I thought it was a good idea that we could try.
After lunch we had some discussions about generative AI, first as a tool to create teaching materials, and then about what we are all doing to educate students on the pros and cons, the do’s and do not’s. James from Middlesbrough talked about their policies and showed us the ‘Acceptable-Unacceptable’ scale graphic that they have been widely publicising. Something else I really liked the idea of.
Finally, we had a roundtable discussion on the state of the VLE as a concept, stimulated by Anthology’s financial woes. Those in the know were very confident that Blackboard isn’t going anywhere, and will be spun off as it’s own thing again. I asked a question about market disruptors, to see if anyone knew of any providers stepping into the market and trying to shake things up a little, the way Instructure did a decade ago. Not a lot there, but I did learn about a new Moodle development specifically aimed at the HE sector which aims to streamline the app and make it more relevant for our community specifically. I had a DuckDuckGo for this, and found this presentation about the project from the Moodle Moot in September which may be of interest to some folks.
AI Disclaimer: There is no ethical use of generative artificial intelligence. The environmental cost is devastating and the technology is built on plagiarised content and stolen art, for the purpose of deskilling, disempowering and replacing the work of real people.
We’ve had a rebrand. Now that we have a new Deputy Vice Chancellor who has changed the title to DVC Education, our annual Teaching and Learning Conference has been renamed to the Education Conference to match.
The day began with welcome messages from the DVC, and our VC, David Bell, who talked about the challenges of identifying truths and falsehoods in our increasingly siloed and partisan culture, and the importance of ensuring students develop critical thinking skills to cope in both education and employment.
The keynote talk was delivered by James Coe, Associate Editor (Research & Innovation), at WonkHE, and a local lad from the North East. The key message of his talk was about the challenges and pressures students now face as a result of cultural changes since his own time at university, back in ye olden days of 2011, before the start of £9k student fees and a time when he received a £4k bursary. Now that students face far harsher financial challenges and graduating into a stagnant labour market, James talked about how pressures have flipped, with students now having to fit in lectures and study around work, rather than the other way round as it was in the recent past and, I would argue, how it should be. This leaves them a lack of space and time to study and benefit from the formative experience of being a student.
The name of the day may have changed, but something that stayed the same was the always excellent student panel discussion. I have always found this very useful and insightful. This year a lot of the discussion was about the real world use of generative AI tools, as might be expected. The panel talked about how they are using these tools to help structure their work and adjust their writing voice, and were well aware of the dangers of overuse and offloading their thinking to these tools. Specifically, they commented about a fear that they would reduce their writing skills. I was also pleased, if that’s the right word, that the panel echoed my concern that the University does not provide sufficient and clear guidance for students on what they are and aren’t allowed to use exactly, and how they are allowed to use them.
I joined this UCISA webinar at almost the last minute, when I found out that they were going to be talking about Cadmus via the HeLF email list. Cadmus is something I need to learn a lot more about in connection with a project I’m involved with this year.
It opened with Julie Voce, of City St George’s, University of London, delving into some of the challenges the sector is facing in relation to generative AI. She talked about human-based detection of AI plagiarism by looking for hallmarks of LLM content, such as the use of words like ‘delve’ and the em dash—for an explanation of why this is problematic, watch Etymology Nerd’s short on the feedback loops which I’ve embedded above—Julie also talked about a practice she has observed in staff, finding that some people will dock students a few percentage points when marking if they suspect LLMs have been used, but can’t prove it.
This led on to Tom Hey’s case study on their use of Turnitin’s AI detection tool at Leeds Beckett. They have been a longtime user of Turnitin’s authorship tool which launched in 2019 to help detect contract cheating, and adopted the AI detection tool when that launched as it was already part of their license, academics wanted it, and they didn’t want staff using unauthorised tools at their own discretion. Tom reported good success with this, but noted that it had to be framed by their ‘Academic Honesty Policy’ for staff and students, which emphasises that these tools are a backstop to help academics, and are nor foolproof detectors of plagiarism. In the chat, someone posted a link to a Jisc paper on the validity of AI detection systems which makes for interesting / depressing reading (delete as appropriate).
Finally, Chie Adachi from Queen Mary University presented about their experience of using Camdus to support assessment during a pilot which ran over the previous year. Unfortunately I didn’t get to see the tool itself or learn much about it, but the results of the pilot were very positive, with 82% of students reporting a positive experience, a 7% increase in average grade, and a 38% decrease in first time failure rate.
Another couple of useful links from the discussion which I thought worth sharing: A Harvard Business Review article on how ‘workslop’ is harming office productivity, unfortunately behind a paywall, but you may have access through your Library, and a report from MIT (PDF), on the impact of Generative AI on business to date – “95% of organizations are getting zero return”.
AI Disclaimer: There is no ethical use of generative artificial intelligence. The environmental cost is devastating and the technology is built on plagiarised content and stolen art, for the purpose of deskilling, disempowering and replacing the work of real people.
