Darren Leslie

@dnleslie.bsky.social

Teacher writing about teaching, learning, leadership and evidence-informed practice.

Why do some pupils avoid answering, even when they may know? Mistakes are one of the clearest windows into pupils’ thinking. Good classrooms welcome them, make thinking visible and give teachers the chance to respond before misconceptions become secure.

AI should not remove the thinking through which learning happens. For novice learners, be cautious about tools that generate answers before pupils have built the knowledge to understand and evaluate them. The most promising AI keeps the cognitive work with the learner.

AI can help a pupil finish the task while quietly changing the thinking the task was meant to develop. Does it replace retrieval, explanation or problem-solving? Or does it strengthen them? The real value of AI lies in the thinking it leaves with the learner.

I want to deepen my understanding of AI and learning. What are the strongest studies or papers on AI being used to teach, tutor or support pupils? I’m particularly interested in research that examines learning outcomes, not only engagement or efficiency. Please share anything I should read

The timing of AI support may matter as much as the tool itself. Before a pupil attempts the work, AI may remove the need to retrieve, reason or persist. After an attempt, it may expose an error, offer a useful hint or improve the next response. Same tool. Different moment. Different effect.

Why do some pupils seem to understand everything in the lesson, then remember almost nothing a week later? Performance during a lesson is not the same as durable learning. The real test is what pupils can retrieve, apply and connect once support has gone.

AI can make a pupil’s work look better without making the pupil any better. That is the distinction schools cannot afford to miss. A polished answer is not proof of learning. The real test is what the pupil can still understand, remember and do when the AI is removed.

We often assume that struggle is evidence of learning. But productive struggle asks pupils to retrieve, connect and apply knowledge. Unproductive struggle comes from confusion, overload or missing prior knowledge. Challenge should deepen thinking, not block access to the content.

AI is at its most useful when it supports thinking, not when it replaces it. Learning is built through retrieving, generating, connecting and refining ideas. So the key question is not: “Can AI do this task?” It is: “What thinking do I want the learner to do?”

Teacher talk is often treated as something to minimise. But reducing it too far can leave pupils with less knowledge, less clarity and less capacity to work independently. The goal is not less explanation. It is better explanation, given at the right moment.

1/ A valuable article from David Didau. Schools search for visible signs of learning because learning itself is invisible. The problem begins when those signs become targets, and teachers start producing the appearance of learning rather than learning itself. 🧵

David Didau@didau.bsky.social · 4w ago

Is Rob Coe’s idea of ‘poor proxies for learning’ “one of the most damaging ideas in education”? Or is it an essential language to help us avoid being seduced by nonsense? Find out below… open.substack.com/pub/daviddid...

If AI gives every pupil instant feedback, will they learn more? Not necessarily. Feedback is one of the most powerful influences on learning. But only if pupils think about it. 🧵

Differentiation is often presented as creating something different for every pupil. But could inclusive teaching sometimes begin with a stronger shared classroom experience instead? 🧵

Can AI think for a pupil? It can generate an answer, explain a concept and suggest an example. But learning does not happen because information exists. It happens because pupils think about that information. 🧵

✈️ Evidence-Informed Wisdom is here!! ✈️ RT for a chance to be the first to read it. Today we received our advance copies. It’s a serious book written in a funny way. It’s my favourite thing I’ve ever done. Will pick someone at random on Sunday night and post them a copy!

Bild

There are lots of ways to provide opportunities for retrieval practice. There are benefits/limitations to different techniques that's why I embrace variety. As a history teacher, quizzes are great for knowledge recall but students must be given opportunities for elaboration & application too.

Bild

I might be wrong but could it be because we simply don’t have a shared understanding of what excellent looks likes Which then means excellent isn’t shared explicitly in a way that schools can look at and work towards. Do we all know (and agree) on what excellent looks like?

Henry Hepburn@henryhepburn.bsky.social · last yr.

ICYMI: Top inspection ratings out of reach for Scottish schools www.tes.com/magazine/ana... @emmaseith.bsky.social @tesmagazine.bsky.social

Paul gave a great summary of the decades of change and increasing pressure they have put on teachers, coupled with the diminished supports available. Really struck by the idea of a morning academic timetable and afternoons for sports/creativity. It'd be great to share what I love beyond my subject!