The Rise of The Machine

I'm scared of technology.

There, I said it.

As a kid, I wasn't. I was using MS-DOS prompts at 6 or 7. I was editing config.sys files by 10, I built a few PCs in my teens, I embraced MP3s, Napster, BitTorrent, ICQ in their early years. I loved tech, I loved new tech, I loved experimenting.

Now... I'm terrified. I don't know where the tipping point was - I was okay with the move to MS Teams during Covid, I bought a graphics tablet and started a YouTube channel making silly illustrated children's books... I even started to make this website last year! But, somewhere around turning 40, my brain just stopped being able to keep up with newness. My inclination to experiment, to try new things, to see what tech could offer me... it disappeared, replaced with a yearning to stick to what I know.

The problem is, the world didn't stop. AI isn't stopping. Far from it. And, in my move to a new world with my new school, part of the Duke's Education group, I have landed in a department who absolutely embrace technology. Paperless teaching, OneNote, Google Classroom, a whole suite of apps and websites designed to support every facet of the child's learning - even experimentation with AI-based marking.

I make a mean excel spreadsheet (and can calculate the mean with an excel spreadsheet). I'd go as far as to say I 'excel' (sorry not sorry) with much of the range of MS Office, as do many of my millenial brethren. But, it is a gigantic leap from this to embracing AI support in all facets of teaching. Where to begin?

So, that is where I'm at. Where to begin? Over the next few months, this blog will mainly be focused on how I am adapting to the challenges of EdTech in 2025. I think I've gotten my head around OneNote to use it functionally, and I've learnt on Google Classroom how to schedule and reuse posts to save time. But AI... save for occasionally checking my workbook answers with it, that is a new frontier, and one I will overcome. I simply must.

My first dabbling, however, was a setback. My HoD suggested using it to mark some A-Level assessments, and then comparing the findings with my own marking. I'm an examiner, so I'd expect my marking in general to be reasonably accurate to within perhaps 1% of the true score. I'd been given figures of purported accuracy to within perhaps 5% with AI (Gemini was the tool of choice). So, I gave it the brief, I uploaded the MS and asked it how it would apply each mark - which it explained with reasonably good accuracy (save for calling 'B' marks "bonus" marks, which elicited much laughter from yours truly). So, without further ado I uploaded a pdf containing all the scanned scripts and set it off. It produced a question-by-question score for all the students, and wrote some convincing-looking feedback (though after the first three it stated "Similar detailed feedback continues for the next 9 students" which brought the second huge chuckle.)

However, my heart sank when looking at the first comparative data produced once I marked the scripts myself. The first one I marked showed ten mark differences (the paper was out of 29, so that is a HUGE error), the third another 9. Even a few of the scripts where it managed to give the same overall score, there were differences in where it awarded the marks. It gave one student full marks on a proof question - I assume because at each stage they were equating the LHS with the given RHS result - though they got nowhere near completing the proof. Overall, the marking was inaccurate by significantly more than 10% - useful as nothing more than a curio. Perhaps a confidence boost that I'm unlikely to be out of work as an examiner any time soon.

But, because this is the year I embrace AI, I cannot just give up. I mustn't. I shalln't. This week, when time allows, I am going to repeat the experiment with another group. This time, however, I will probe the AI further before scanning in scripts. I will check that it really 'understands' some of the M marks. I will embrace its failings as my failings, and I will seek to be a better teacher to it. Perhaps then it may be more useful to myself and my students.

Whatever happens, for better or worse, I'll keep you updated.