The antidote to anxiety is action

Assessment is such a hot topic, now and for my entire career. But I have rarely sensed such tones of urgency, despair and anxiety surrounding it since the arrival of AI in education. It can drain all the oxygen out of the room in any training session I do, if I allow it. And it has been difficult to support over the past few years as there were few hard and fast solutions to offer to schools in need.

I have to admit I have shared those concerns every step of the way with schools. There were various times in 2024 and 2025 when I would have bet my house on Internal Assessments becoming extinct, so AI vulnerable were they. The alternative of a single high-stakes assessment on a single day did not please me either but I could see no way to verify or validate work while the detectors were so unreliable. The only session, ironically where I did not feel this urgency was in a Nuffield Assessment research meeting where all the major awarding bodies seemed quite sanguine. They seemed fairly confident (whether due to professional pride, competitiveness or otherwise) that a solution would come and / or that schools were best placed to authenticate student work.

As such, many schools have been patching, bridging, experimenting and hoping for the best with assessments over the past few years. I had precious few great examples of success upon which to draw but could see lots of trends and patterns occurring:

Eblana Learning has a bias to action, looking for practical supports and cautious optimism, where reasonable and researched. This has meant that in sessions, I tend to share things that are working in pockets of great practice around the world and encourage them to continue the things we always do: know your children and their voice and style as well as you can, understand the nature and the need of your assessment, experiment, adjust at the edges, innovate where we can, think about our students, think about valid assessments, and rethink assessment opportunities.

Work within the constraints and the parameters we have. I always remind them that if you think we have it bad, the universities are really struggling even worse than we are. We are also minded of the time between Google/ Wikipedia and the advent of Turnitin.com. We struggled. We were frustrated, but ultimately there was a technological and a pedagogic solution to that copy-and-paste world, and I believe there will be for this too.

Conversely I have become tired of the platitudes around assessment conversations online:

“we need to rethink assessment!” - great, please tell us how! The next sentence giving detail is often absent…

‘viva voce or verbal defence” - true, a good idea but if you’ve ever tried to give one-to-one feedback to a whole class for IAs or schedule oral exams, imagine the scale of disruption to this for every subject, in every year for each assessment… where would we find the time to do some actual teaching and learning…?

“process over product” - also nice to say and hear but what does it mean for a solutions-based subject? And as one assessment colleague told me, as soon as you assess the process, the process becomes the product. And what if the process becomes the product?

It reminds me of the days when I would join Weight Watchers, and on that first week I would make sure to have drunk as much water as possible, and sometimes even put rocks in my pockets to make sure that my first week loss was significant. And so students, or at least those who value results over learning, will always find a way to gamify the process, invalidating the assessment.

Every so often, a model catches my eye and optimism returns. I hear a lot of chatter about the Swiss cheese model. The idea is a “multi-layered defense strategy that stacks imperfect methods—like assignment design and oral checks—because no single tool can stop academic dishonesty or learning loss. [1] (https://gptzero.me/news/how-should-educators-approach-ai-and-academic-integrity/)” Although originally a risk management method for industry from the 90s, it was applied to assessment by Dr. Kiata Rundle and first came to my attention thanks to Professor Phillip Dawson (Deakin University) via the AI in Education Podcast.

I worry that so many assessments will automatically lead to student fatigue and teacher burnout and have fantasised about a simpler polaroid model where two perpendicularly design assessment somehow blocked everything out. I haven’t figured that part out yet but will keep trying. If you don’t remember it, the polaroid experiment is beloved of science teachers (me!) to show how it allows light (cheating) in one alignment but blocks out light completely (AI-proofing) when the orientation changes.

Over the last year I have thought so much about those at the front line of assessment - like the teacher who dejectedly put her head on the desk and said, “I feel like every day is a battle against AI in my classroom, and I’m losing badly.” Or the Coordinators and HODs who tell me that their lives have become full-time AI police or it has introduced impossible friction into their relationships with students and parents. I longed to give them hope.

Having just been in India at an incredible school - Dhirubhai Ambani International School- I saw those pieces of advice coming together in a way that genuinely excited me. They had taken many of the things we talk about and actually put them into practice: regular check-ins, verbal defences, high expectations, strict management, regular reinforcement of AI ethical considerations, clear disclosures. The result was a coherence and a calm that you feel the moment these teachers spoke about their classrooms.

At Dhirubhai Ambani International School, they spoke about some very deliberate choices they are making, choices which I neither endorse nor refute but acknowledge here fully. These choices were things like tying their predicted grades to their internal assessment quality and finding a better word for “compliance” that emphasised how important these were.

What the teachers articulated so clearly was how their approach released pressure. Yes, it ensured that students front loaded a great deal of work into the IA, but it released so much pressure with the teacher-student frictional relationship, with deadlines and chasing and so on. They described that perfect scenario that I’ve always wanted as a DP teacher, where the IAs were done and dusted, and you could really focus on preparing for external assessments. This year, their results were simply incredible.

They painted for me that perfect scenario that I’ve always wanted as a Diploma Programme (DP) teacher, where the IAs were done and dusted in every subject early in Year 2 and you could really focus on preparing for external assessments. In fact, this year the results were simply incredible. That’s their story to tell. All I will say is that I have only ever dreamed of having an entire class fully focused, with a rich and interesting (and authentic!) internal assessment in the bank. They really are living the dream. When I said that to them, I could see the realization dawning on them, and I said, “This is fantastic. We need to share it with the world.” Hence this article.

One swallow does not a summer make, and they acknowledged this very clearly themselves. If this is something that we see repeated over time, I think this is something we really, truly need to capture and extend and share widely with our incredible, kind, sharing IB community. They promised to continue to track the data to adjust and refine the approach, and I’ll report back next year as we wait for seismic shifts in assessment.

Change is possible, positive and, as the models change, so too will teachers adapt and innovate… and we’re going to be okay.