Teach Secondary -Issue 15.1

becoming dependent upon it. The DfE has also helped set the tone through its own guidance on use of AI in schools. The government message is that teachers can use AI to help with planning, resource creation, marking and admin, but that all professional judgements and responsibilities must sit firmly with the human adults in the room. In the DfE’s view, teachers are ‘irreplaceable’, with the main aim for AI being to free up teachers’ time for more face-to-face work with pupils. Use of AI by students is treated more cautiously. Schools are reminded that many mainstreamAI tools are for use by over-18s only, and that any pupil use of AI will require appropriate safeguards, filtering and supervision, alongside careful consideration of AI’s implications for learning and homework. Showing yourworking I’m proud of being a science teacher. We teach students to think. Science isn’t really about electromagnetism, electrolysis or photosynthesis; it’s about howwe gather evidence and howwe use it, and AI is an awesome tool for assisting with that. The Review explicitly acknowledges this when it talks about ‘media literacy’, pointing out that subjects such as history and science already help young people to analyse claims, weigh up evidence and spot bias, and in the case of science, critically and empirically evaluate various scientific claims. All this matters muchmore in an age when realistic photos, videos and audio can be generated in seconds by anyone with a web browser. Personally, I love using tech inmy classroom. In the past, I’d record students’ results during an experiment and quickly use Excel to plot a graph, which we’d then analyse together on the board. This form of rapid analysis was something I wanted them to be able to do over the next few lessons. Now, I simply snap a photo of the students’ results, and AI makes a graph showing us the relationship in what they’re doing within just a few seconds! But, youmay ask, what’s so aspirational about that? Lots , I would argue, in that it allows us to do more, better, in less time. Analysing the steps taken by an AI to reach an answer is just as useful as peer marking, or using a model example of what a good answer looks like. And if it produces an error, re-prompting the AI to correct its working is precisely what skilled users of AI are already doing to create output of hitherto unimaginable quality. Conversation and experience Where I see AI mostly improving education is in giving us a chance to put human conversation and practical experience back at the very centre of what we do. In a good science lesson, the most important moments will often be those when you’re stood at the board, arguing with the class about what a graph really shows, or walking around the room as they wrestle with a tricky practical and gradually find their way through it. If an AI system can quietly generate differentiated questions, draft model answers or provide individual support when students are stuck, that leaves me free to spend more of my time in those moments. The technology becomes a background tool that helps us think together; not a replacement for that thinking. When I imagine the classroom 10 years from now, I don’t see myself walking in and opening a PowerPoint file. I imagine systems that already know what my department has planned for the term, which have analysed prior assessments, and can suggest starter questions matched to where each student actually is that point in time. As the pupils file in, AI might help me choose a calming image to display on a Friday afternoon, or a striking graph for a Tuesday morning, since the system will have learned what helps the students to settle and readily engage. While they work, AI tools could be proposing individual tasks, drafting plenaries and setting tailored homework that’s sent direct to their devices, while I move around the roomwatching, listening and guiding the thinking. If we get this right, the next version of our curriculumwon’t ask pupils to pretend AI doesn’t exist. It will assume that they’ll be maturing in a world where models can scan, summarise and simulate at speed, and focus instead on helping them ask better questions, understand data and evidence, and knowwhen not to trust the first answer on a screen. Our job will be to make sure that the generation inheriting these tools can still notice what’s missing, hit CTRL+F and insist on seeing the whole picture. ABOUT THE AUTHOR Kit Betts-Masters is a lead practitioner for science and produces physics, education and technology videos for YouTube under the username @KitBetts-Masters; for more information, visit evaluateeverything.co.uk 53 teachwire.net/secondary E D T E C H