Mar 26' Update from Good Future Foundation
GenAI and Functioning Systems - What This Means for Schools
The conversation around AI in education is still catching up with what the technology can now actually do. Not long ago, most people associated generative AI with producing text, images, or short pieces of media. That framing is already outdated because the more recent shift is towards systems that can be designed, iterated and deployed to perform structured tasks with utility. In practical terms, this means AI tools enable even the most modestly experienced user to build something impactful.
This has implications for how young people engage with what they are learning at school. When used carefully, these systems can create new entry points for students who might otherwise disengage. The ability to respond dynamically, to adapt to different needs and to present information in varied formats offers a level of accessibility that traditional approaches often struggle to achieve. For some learners, particularly those who require more tailored support, this could materially improve how they develop understanding and confidence over time.
At the same time, the reduction in technical barriers introduces a different category of risk because it is now entirely feasible to assemble functional tools, including conversational systems, without a clear understanding of data protection, safeguarding, or failure modes. Schools will soon be dealing with an emerging layer of digital capability that can influence communication, decision-making and the handling of sensitive information.
Spending time building and testing these systems makes the trade-offs more visible. The quality of what is produced depends less on the act of 'using' AI and more on the clarity of intent, the structure of the system and the constraints applied to it. Without those, outputs become unreliable quickly, but with them, the same tools become significantly more effective and more consequential.
For schools and education leaders, the priority is to develop as much informed judgement as possible. Understanding where these systems add value, where they introduce risk and how they should be governed is becoming a core part of responsible practice. This is the space the Good Future Foundation is working in, and we remain here to support you in your onward journey.
Conversations on AI in Education at Next Generation Schools Conference
The Good Future Foundation was proud to support the Next Generation Schools Conference 2026 on 6 March and celebrate the national rollout of the Big AI Project!
Before the conference officially began, key partners of the Big AI Project gathered for a roundtable discussion where our Executive Director, Daniel, facilitated conversations around key issues facing education in an AI-infused world, the foundations needed to support schools and partners, and concrete next steps for collective action.
Daniel and our Project Manager, Anneliese, then led a breakout session talking about how the AI Quality Mark can help schools build safe, thoughtful and strategic approaches to AI. The conference concluded with an on-stage panel discussion where Daniel chaired “Learning or losing control in the age of AI” featuring three excellent panellists: Rebecca Mace, a senior leader in higher education and AI-in-education research; Jade Bowler, an educator and creator focused on youth empowerment and system reform; and Ali Gellett, the Big AI Project Lead.
Sharing our work with over 200 passionate educators was energising. These weren’t surface-level conversations. Educators and school leaders were wrestling with what thoughtful AI adoption in schools should look like.
If you’re a school leader looking for more support, you can now book the training for free and gain access to teacher training online sessions and a suite of AI literacy curricula. We look forward to seeing how schools across the UK engage with the Big AI Project!
Bringing AI Professional Development Across the UK
Early March perfectly embodied our mission at Good Future Foundation in making sure tech-agnostic professional development and AI guidance reaches educators everywhere, not just in specific regions.
Our team was on the road from Cornwall to Berkshire, running in-person sessions across multiple locations. We delivered workshops in Cornwall for the Kernow Learning Trust’s Inset Day, while also running sessions in the West Midlands with E-ACT on the same day.
Schools also came together at The Royal Grammar School, High Wycombe for our regional event on responsible AI strategy and culture. The sessions covered choosing the right tools for specific tasks and thinking through purposeful deployment.
The month’s events culminated with the Eton Star Partnership conference at Dorney Lake, where Daniel joined a panel discussion on inclusive practice alongside fellow education leaders and practitioners: Dr Fiona Aubrey-Smith, Tom Wade, Dr Stella Scharinger and Peter Reeves.
These events reinforced the immense value of in-person engagement. When school communities gather around specific themes, it builds momentum for ongoing dialogue and exploration. And our support for schools remains completely free, reflecting our commitment to accessible AI guidance for all educators.
Looking ahead, we’ll be collaborating with Xavier Catholic Education Trust in April for another professional development event. Please find out more and register below.
When Teachers Build Their Own AI Tools at Haberdashers' Elstree
Haberdashers’ Elstree Schools (Habs) comprise two highly academic 4-18 independent day schools, educating over 2500 students. We were delighted to have recently been recognised with the Gold AI Quality Mark by the Good Future Foundation, and would strongly encourage other schools to engage with the GFF. We are immensely impressed with the deep understanding of education that shapes the organisation, and its firm grounding in ethics and the best interests of children.
Below are a few illustrative examples of how AI is shaping everyday experiences of teaching and learning at our schools. Our teachers, and students in Year 9+, have access to Copilot, Gemini and NotebookLM secure within our tenancy, and we have just given teachers access to Google AI Studio, giving them better opportunities to build their own apps.
Engaging with Shakespeare
The English department in our Girls’ School has been exploring how AI can enhance and enrich everyday classroom learning. One of their recent experiments was a Shakespeare murder mystery activity, where AI helped bring the task to life: the teachers used it to generate realistic, character portraits in the style of The Traitors, and create a simple interactive page where each portrait opened into a detailed case file. The mix of striking visuals and the easy click-through interactivity made the experience feel more authentic, helping students engage deeply with motive, evidence, and close textual analysis.
Across the wider English curriculum, AI is used in small but meaningful ways to harness the engagement and accessibility of digital learning. Without any formal coding knowledge, the department has been able to vibe-code digital card sorts – saving a huge amount of cutting-and-sticking time – and built quick interactive games that reinforce key terminology and give students instant feedback. Because these tasks are on a digital platform, they feel genuinely low-stakes: students feel more comfortable experimenting, trying, failing, and simply pressing “undo” as they refine their understanding. For KS4 and KS5 classes, they added export functions so students can save their sorted tables or digital annotations straight into their notes.
Transforming History Teaching
Our Boys’ School History department has integrated AI across teaching and learning in ways that have significantly enhanced both efficiency and impact. AI is used to scaffold complex assessment questions—such as GCSE change and continuity tasks on medieval medicine - and to generate word banks, glossaries, and role‑plays, for example on the abolition of slavery. It supports students’ essay planning by extracting precise evidence from academic books such as Cannadine’s Victorious Century. AI can also be used to create model answers on topics such as the development of warfare, produces revision materials and flashcards from exam questions, and generates tailored writing frames to support extended writing. The impact of these tools has been substantial: AI has made it simpler and quicker to adapt our teaching to meet learners’ needs, allowing us to personalise support, strengthen historical writing, and redirect teacher workload.
Building sixth formers independence
At Sixth Form, where students from the Boys’ and Girls’ Schools learn together, AI empowers independent learning and metacognition. In French, for example, using Copilot on scanned handwritten essays, teachers highlight error categories, such as grammar or idiomatic usage, without providing any direct corrections. This data feeds into a colour-coded OneNote starter activity where students are challenged to correct their own mistakes.
Students also utilise “Learning Accelerators” such as Speaker Progress for independent speech recording, receiving real-time coaching on pace, filler words, and pronunciation. Across the broader Sixth Form, tools like NotebookLM act as secure personal assistants, generating bespoke study guides and exam questions directly from students’ own class notes to avoid the hallucinations common on other platforms.
There are also many cases where AI becomes part of the academic focus for the lesson itself. Following a recent viral post in which it was claimed that AI created a vaccine for a dog, Biology students studying mRNA vaccines assessed the veracity of this claim, the role AI really played, applying their scientific knowledge, and understanding that one case does not constitute a robust scientific finding.
The keys to effective integration
Two things about our teachers’ approach are inspirational. First, every use supports academic endeavour and never takes a cognitive shortcut. We are fuelled by love of learning, and that is manifest in their approach. Second, because we are so lucky to have such intellectually engaged staff, they have risen to the opportunity to build bespoke activities and apps to use with students. Our pioneer teachers, who started vibe-coding nearly two years ago, have led the way, and now it is becoming part of many teachers’ toolkits. By putting development tools directly into the hands of our educators, we ensure our AI integration remains firmly grounded in the realities of the classroom.
AI Quality Mark Provides Schools the Framework to Embed AI Effectively and Responsibly
We recently commissioned an independent impact evaluation to understand how well we're serving schools through the AI Quality Mark. The findings have been both humbling and encouraging, and helped us see where we can continue to grow alongside the schools we support.
To read the full report, please click here.
Our AI Quality Mark framework is completely free. More and more schools are already part of this community, and we’d be honoured to support you too. Please reach out to us to receive detailed information and book your initial conversation.
New Awarded Schools
We are delighted to celebrate 17 organisations that have recently been awarded the AI Quality Mark! These include schools in UAE, Spain and Thailand, one school that progressed from Bronze to Silver in just 9 months, and three school groups and multi-academy trusts representing 25 schools in total!
Gold Award: GEMS Cambridge International Sharjah, Berkhamsted Schools Group, St John the Baptist School
Silver Award: Cranleigh School, Beech Hall School, St Albans School, Thames Christian School
Bronze Award: Saint Catherine of Siena Catholic Primary School, Acorns Primary School, Epping Forest Schools Partnership, Shrewsbury International School Riverside, Chipping Ongar Primary School, The Solent Schools, West Leigh Junior School
Progress Award: The Complete Works, British School Alzira, Xativa & Gandia, Dulwich College
Many of these schools are sharing their journeys through our newsletter and community platform, providing a lot of valuable resources and insights for other educators who are also navigating AI use. We also host bi-monthly virtual community meetings where teachers discuss their experiences trailing AI tools and ask questions. Our next meeting is scheduled for 4pm on 22nd April, Wednesday. We look forward to seeing you there!
Park School: How AI Is Enhancing Learning, Teaching and Wellbeing
This month, we have the pleasure of inviting Mrs Bernie Davis, Principal of Park School in Belfast which recently awarded Bronze of our AI Quality Mark, to share how the school is exploring the use of AI to support their students with special educational needs.
Mrs Davis has dedicated 30 years to working and special educational needs, building her career from the ground up with determination, humility and hard work. She began her journey in education as a classroom assistant, later qualifying as a teacher. She went on to serve as Key Stage 3 head and Vice principal and has proudly led Park School as principal for 11 years. Her leadership is defined by compassion, high expectations, and an unwavering belief in the potential of every child. She worked tirelessly to create a nurturing, ambitious and inclusive environment where pupils with special educational needs are valued, supported, and empowered to thrive.
AI Is Not a Subject
In Short
Jensen Huang says every graduate should be an AI expert. Three independent studies published this month show why that is urgent — and what “expert” should actually mean.
Most AI education efforts treat AI as a subject to study. Understanding how models work is useful; learning to work with AI to elevate your thinking is essential.
The skills that survive AI disruption are not technical. They are judgment, verification, synthesis, and communication.
For educators: the question is not whether to teach AI, but whether we are teaching the right things about it.
The question
NVIDIA CEO Jensen Huang told MIT professor (and podcaster) Lex Fridman this week that every college student should graduate an AI expert, and that he would always choose the AI-proficient candidate. “A carpenter with AI is also an architect,” he said. “Their artistry just elevated tremendously.”
That sounds right. But what does “AI expert” actually mean? The education world is trying to answer that question. The AI4K12 initiative has built the dominant US framework for teaching AI in schools, organised around Big Ideas like perception, reasoning, and machine learning. UNESCO published an AI Competency Framework for students in 2024, covering AI techniques, applications, and system design. Meanwhile, prompt engineering courses are drawing hundreds of thousands of enrolments online.
These are valuable initiatives, but they share a common approach: AI as a subject in its own right — something to be studied, understood, and mastered as a technical discipline. But is that enough? The labour market is telling a different story.
This month, OpenAI co-founder Andrej Karpathy released a visualisation tool (above) scoring 342 US occupations on AI exposure. The single strongest predictor of a high score: does your work output live on a screen? 130 occupations scored 7 or above, representing 49 million jobs and $3.7 trillion in annual wages — overwhelmingly knowledge-economy roles like software developers, financial analysts, writers, and graphic designers. Workers earning over $100,000 average 6.7 on exposure. Those under $35,000 average 3.4. The economic hierarchy that placed digital knowledge work above manual labour is being inverted by the very technology that created it.
Anthropic’s labor market study tells a similar story from the usage side — measuring not what AI could theoretically do, but what it is doing, using real-world data from Claude. Computer programmers: 74.5% of tasks already covered. Customer service representatives: 70.1%. Unemployment has not spiked — but hiring of workers aged 22–25 into exposed occupations has dropped roughly 14%. These are the students we taught four to six years ago.
Both studies agree on what is being automated — first drafts, data processing, routine analysis, code generation — and what remains stubbornly human: judgment, verification, synthesis, and communication. Most AI curricula teach students to understand and operate the technology. In industry — and I see this daily in my own role at a tech company — nobody learns to use AI by studying how models work. They learn by using it to solve problems they already care about, and the skill is not understanding the tool but thinking clearly enough to direct it.
That does not mean technical understanding is worthless — it is genuinely valuable for students who later specialise in computer science, or at a postgraduate level where the plumbing matters. But it is not where you start. You learn to drive by driving: road rules, parking lots, building up to motorways. Nobody thinks the right starting point is internal combustion. If you decide to become an automotive engineer, by all means learn how engines work. But that is specialisation, not literacy.
What I teach instead
When I walk into a classroom, I do not start by talking about AI. I start with a problem worth solving.
In an interdisciplinary course I teach at the Hong Kong Academy for Gifted Education, the challenge is this: Hong Kong has heritage brands — small, family-run businesses with decades of history — that are struggling to survive. Their marketing is stuck in an older style: selling features rather than telling stories, weak social media, websites that have not been updated in years. These are businesses worth preserving. I show the students the problem and ask them to help.
AI never appears in the brief, but it becomes the means for everything that follows. Students use AI to research the brands, analyse competitors, generate campaign concepts, and draft slogans, stories, visual ideas. A lecture about how language models work would not have helped. A problem worth caring about did — and AI became the way to engage with the topic at a depth that would otherwise have taken weeks.
I see the same dynamic in my day job at a tech company. I do not teach colleagues about neural networks. I help them use AI on the problems they already have — teaching workflows, processes, and mindsets that transfer across tools; values comes from knowing how to use AI to think and work at a higher level.
In effect, I organise what I teach around four competencies.
Information literacy — finding, assessing, and synthesising sources with AI, then evaluating what it gives you. In that same HKAGE class, students had to fact-check each other’s AI-sourced claims about the brands. Facts and statistics regularly turn out to be fabricated. That is not a failure of the lesson. It is the lesson.
Analytical planning — using AI as a thinking partner rather than an answer machine. The students who produce the strongest work are the ones who don’t accept AI’s first response. They push back — asking “what am I missing?”, challenging assumptions, iterating until the plan fits the context. The difference between using AI well and using it poorly is whether you treat it as a sparring partner or an oracle.
Technical creation — this is where practical AI skills come in: using AI to write code, generate images, draft copy, and create prototypes. The first version is never right, and the cycle of writing, testing, and refining is where real learning lives. I wrote about this last month with Essay Hero: the AI wrote the code, but every decision about what to build, when to stop, and whether it actually worked was mine.
Critical communication — presenting and defending your work to others, evaluating claims, giving structured feedback. Of the four, this maps most directly to what the research says survives. Relationship-heavy, trust-dependent, communication-intensive work scores lowest on every AI exposure index. A student who can explain their choices, challenge someone else’s reasoning, and defend a position under questioning has skills no model can replicate.
None of this requires understanding attention mechanisms or memorising prompt templates. It requires thinking clearly, verifying carefully, building iteratively, and communicating what you know. The tools will keep changing. These skills will not.
The evolving teacher
If the student’s role shifts from producer to director, the teacher’s role shifts in parallel. When AI can deliver content, generate practice questions, and draft feedback, the teacher’s irreplaceable value moves from delivery to design — crafting the right inquiry for the right student at the right moment.
That means designing activities where the student’s own thinking is irreplaceable — if a student can complete the task by copy-pasting AI’s first response, the activity needs rethinking. It means coaching focus, scope, and prioritisation. And teaching structured critique as a discipline, not an afterthought.
Many teachers do not yet have the freedom to teach this way — current curricula and assessment frameworks were not designed for an AI-augmented classroom, and they constrain the very task design that would serve students best. But if current labour market trends continue, and AI’s impact becomes impossible to ignore, decision-makers will have to update examinations and curricula (at least, that’s my hope). The question is whether schools wait for that mandate or start preparing now.
Huang said a carpenter with AI is also an architect. That is true — but only if the carpenter learned to think like an architect, to evaluate designs, make trade-offs, and take responsibility for what gets built. AI tools are more accessible than ever, and what costs hundreds of dollars a month today will be free in a year or two — that is just how computing works. The question is whether we are teaching the thinking that makes those tools transformative, or just the tools themselves.
If you want to try one thing this term: take a lesson you already teach, and redesign one task so that the AI does the first draft and the student’s job is to evaluate, improve, and defend it. That is where the learning lives now.
Sharing Students Perspective with UnJaded Jade
Earlier this year, our Student Council met with Jade Bowler, an advocate working at the intersection of youth empowerment, system reform, and mindful education, to share the AI-related issues they’ve been discussing and exploring as a Council.
Jade later created a video called “Every Student Needs to Watch This Before Studying with AI” for her Youtube Channel, UnJaded Jade, which has nearly one million subscribers - many of them are students who follow her study tips and approach to making learning enjoyable. Some of your students might already know her work.
Feel free to share this video with your students to give them a peer perspective on using AI thoughtfully and responsibly.
Celebration of Best Practices of Responsible AI Use Among Students
In collaboration with Queen Ethelburga’s Collegiate, our brilliant Student Council has been planning for months to launch the Smart & Safe AI Use Student Initiative. To make this fully accessible and inclusive to all students regardless of location, we’ve transformed it from a traditional competition into a celebration of best practice in responsible AI use.
The Student Council just delivered the first workshop on responsible AI use in late March, and we’re very happy to see students in UK, UAE and Argentina participating in this initiative! If your school would like to engage your students in this meaningful opportunity, please register below to receive the full information pack!
New Episode on Foundational Impact Podcast: Language as our Defining Asset
Don't miss our recent podcast episode where Daniel Emmerson and linguist Dr. Biljana Scott had a captivating conversation about the intricate relationship between language and artificial intelligence.
Their discussion illuminates whether our increasing reliance on AI might reshape how we think and express ourselves, unpacks linguistic concepts like “presuppositions” in everyday speech, and reveals how the terminology we use to describe AI carries powerful connotations that fundamentally shape our relationship with technology.
As their conversation wraps up, Biljana encourages us as educators to engage students in critical thinking about AI and help them identify the uniquely human aspects of communication that AI simply cannot replicate.
You can now listen to the complete episode across all major podcast platforms.