Alison "Ali" Gellett leads The Big AI Project at the Oasis Big Education Institute. The project aims to give every school in the UK and the Republic of Ireland what it needs to approach AI safely, ethically and strategically. Ali has 30 years of experience in educational leadership across teaching, multi-academy trust (MAT) executive leadership and local government. She is a Fellow of the Royal Society of Arts.

For the past decade, Ali has focused on transforming the education sector by co-designing solutions with schools. She founded and runs the My Future programme, which teaches careers, aspirations and essential skills to primary school children. Since 2021 it has reached more than 35,000 children and enabled over 800,000 encounters between children and employers. It is funded by Essex County Council and has just launched its first London pilot. The Big AI Project grew out of Big Education's Rethinking School programme and is funded by Salesforce. It is free and AI-agnostic, and its materials were designed by teachers and quality-assured by Professor Rose Luckin's team at Educate Ventures Research. After a pilot with more than 230 schools, it is now rolling out nationally. That includes a region-wide partnership with the Liverpool City Region Combined Authority covering its 400–500 schools. Ali co-authored The Big AI Project report (March 2026).

In this episode, Ali talks with Svenia Busson about:

• Why a school's AI strategy should start with vision, purpose and pedagogy, not with choosing tools
• Building a coalition of 50+ partners around shared values, and keeping commercial and political motives out of the room
• A four-hour, AI-agnostic training for heads — and why their confidence rises from 1–2 out of 5 to 4–5 out of 5
• AI literacy from age 3 to 14, with prediction games, bias and stereotyping activities, and class agreements not to share misinformation
• Anthropomorphism, deepfakes and the AI tool that erased the features of a child with Down's syndrome from a nursery resource
• "Truth has gone": misinformation, young people's mental health, and a million 16–25-year-olds not in education, employment or training
• Why human connection and essential skills matter more than ever in a world full of AI
• The Good Future Foundation AI Quality Mark, and why schools that go it alone end up making their job harder

Dr. Ola Ozernov-Palchik is a cognitive neuroscientist who studies how the brain learns to read and, more and more, how AI should (and shouldn't) be used to help it. She is a Research Scientist at MIT's McGovern Institute for Brain Research, a Research Assistant Professor at Boston University Wheelock College of Education & Human Development, and Associate Director of Translational Research in BU's AI & Education Initiative. She got her start in research almost by accident: as an international psychology student, she landed her first research job at the Institute for Reading Research at Southern Methodist University because she spoke Russian.

That job led her to cognitive neuroscience, a PhD in Cognitive Science at Tufts, and early work with Dr. Nadine Gaab at Boston Children's Hospital on some of the first studies to predict dyslexia from kindergarten, followed by a postdoc with John Gabrieli at MIT. She has published 30+ peer-reviewed papers, including a recent Nature Communications study showing the brain's language network is already adult-like by age 4. She received the International Dyslexia Association's Early Career Award, and she is President of New England Research on Dyslexia and Associate Editor of Scientific Studies of Reading.

She founded and directs the Evidence-Based AI in Learning (EVAL) Collaborative, a partnership between BU Wheelock and the Hariri Institute for Computing. EVAL holds AI learning tools to the evidence standard we'd expect of a medical intervention. Its first AI in Learning Challenge drew 35 companies and chose Amira Learning for a free, independent randomized controlled trial. She also co-built KIVA, an open-source AI reading companion that won a 2026 Tools Competition award.

In this episode, Ola talks with Svenia Busson about:

• How the brain learns to read: repurposing circuits for vision and language, the "visual word form area," and why b and d are so hard to tell apart
• What causes dyslexia: phonological processing, family history, and why reading ability sits on a continuum
• Why decades of reading science still don't reach classrooms, with the US as a case study: only ~40% of fourth graders read proficiently on NAEP
• Why no one checks edtech the way medicines are checked: schools buy tools based on user numbers rather than learning outcomes, and some districts are now banning AI outright
• Inside the EVAL Collaborative: pre-registered, open-science RCTs published regardless of outcome, and the first EVAL AI in Learning Challenge
• Can an 8-month trial prove learning? Measurement models, gated technical reviews (including how well speech recognition handles accents and dialects), and evidence that builds up over time
• Advice for edtech founders: collect evidence from day one, and edtech's "5% problem"
• KIVA, skeptical first graders, and the "AI penalty": teachers rated AI tutoring responses higher than human ones until they were told they came from AI

Dr. Nicholas "Nick" Zufelt is Instructor in Mathematics, Statistics, and Computer Science at Phillips Academy Andover in Massachusetts. He was recently appointed Resident Scholar in AI and Education at the school's Tang Institute. He trained as a mathematician, earning a graduate degree and completing a postdoc, then decided research wasn't his path and chose the classroom over a data-science career in tech.

Nick taught himself most of his computer science and machine learning. He now teaches Andover's advanced Machine Learning course, where students build neurons from scratch before touching any library. The day ChatGPT launched, he put it in front of his students and asked, "What do we do with this?" He has been working out the answer with them ever since. As a Tang Fellow, he co-led the ethi{CS} project, which builds ethical deliberation into STEM courses. He is also a founding member and editor of explorations at AI Co-Lab, the teacher-led, peer-to-peer AI professional development network co-founded by Maureen Russo Rodríguez and Nate Green, which has grown to thousands of educators.

In this episode, Nick talks with Svenia Busson about:

• Why computer science still matters when AI writes the code — programming vs. computer science as a liberal art
• Explaining AI to a tenth grader: "code that makes more human-like decisions" — and why "AI" is partly a buzzword
• Metacognition as the core AI skill — the "barometer" (measurement) and the "navigator" (control)
• When and how younger students should meet AI — shared-screen, whole-class use before individual access
• "Practice outside, perform inside, create in both" — a framework for assessment in the age of AI, including oral exams and assessments tightly coupled to learning objectives
• Why teachers shouldn't hand drafting over to AI — lesson plans, parent emails, and protecting your own voice
• Inside the AI Co-Lab model — monthly explorations, intermediary artifacts, and teacher-to-teacher learning
• Two quick wins for teachers: "name the verb" the AI is doing, and the "right-hand side of the chat" reflection prompt

Merlia Shaukath is the Founder and CEO of Madhi Foundation, a Chennai-based nonprofit working to end the foundational learning crisis in India's public schools. She studied Economics at Stella Maris College in Chennai, then earned an MSc in Area Development Studies from the University of Oxford and an MSc in Governance and Management from the London School of Economics. Her grandfather was a social reformer who co-founded educational institutions for India's Muslim community. Her parents were first-generation entrepreneurs.

She was the first staff member at Teach For India, where she helped build the Chennai chapter. She then moved into policy consulting at Athena Infonomics, working directly with state governments. She founded Madhi in 2016 to close the gap between policy and classroom practice. Madhi is the chief management partner of Tamil Nadu's Ennum Ezhuthum foundational learning mission, which reaches about 37,000 primary schools, 2–2.5 million children and around 100,000 teachers. During that work, the share of classrooms getting the target number of teacher-coach visits rose from 1% to 86%, and Tamil Nadu moved up the National Achievement Survey rankings for the first time in a decade. Madhi is now expanding to Meghalaya and Nagaland. Merlia is an Ashoka Fellow, a Mulago Rainer Arnhold Fellow and an Acumen India Fellow, and a Principal Investigator with the What Works Hub for Global Education at Oxford.

In this episode, Merlia talks with Svenia Busson about:

• Why "no one is self-made," and how a social-reformer grandfather and entrepreneur parents shaped a social entrepreneur
• Going from Teach For India classrooms to state-level policy work, and seeing inequity as a systems problem
• Ten years of systems reform: "scale is humbling," scaling programmes back from great to good, and why there's no playbook for implementation
• Why evidence doesn't automatically lead to change, with the politics of multigrade classrooms (70% of Indian primary schools) as the example
• AI in systems where children can't yet read, and why foundational learning has to mean more than literacy and numeracy in an AI-native world
• AI for the adults in the system: lightening the load for teachers and the 3,500 teacher coaches in Tamil Nadu without adding to cognitive decline
• The numbers she cares about most: coach visits rising from 1% to 86%, and "effective teacher" ratings dropping from 90% to 19% as coaches set a higher bar
• Speech-based assessment and oracy, including why Madhi is collecting voice data from 35,000 children to build a Tamil reading-fluency model, and the emotional cost of AI that gets a child wrong
• Why parents are the overlooked lever: 72% didn't know their child was struggling to read, and what non-literate parents can still do at home

Dr Malak Sadek is Course Director and Teaching Associate at the Centre for Human-Inspired Artificial Intelligence (CHIA) at the University of Cambridge, where she teaches the Conversational AI module of the MPhil in Human-Inspired AI. She holds a PhD in Design Engineering from Imperial College London's Dyson School of Design Engineering. Her thesis, Designing Responsible AI: Integrating Human Values into AI Design, received the school's 2026 PhD Outstanding Thesis Award.

Trained as a computer engineer (BSc, American University in Cairo) and in human-computer interaction (MSc, University of St Andrews), Malak created the Value-Sensitive Conversational Agent (VSCA) Framework and Toolkit for co-designing chatbots with the communities they serve. She also co-authored a review of 52 co-designed conversational agent studies (Design Studies, 2023). Her recent paper, Using Anti-Values as Conversational AI Design Constraints (ACM CUI 2026), asks what AI should never do.

She now directs the HIVE CAI Lab at CHIA. The lab studies how people experience values in conversational AI and how they can steer them. Her research is on adults, not children: how adults use chatbots, rely on them and are shaped by them. She is not a child-development specialist. She has worked with The Alan Turing Institute and the Leverhulme Centre for the Future of Intelligence, and was elected Communications Officer of the British Computer Society's Interaction Specialist Group.

In this episode, Malak talks with Svenia Busson about:

• Why bias, accented voice assistants and "secretary" chatbots all come from the same gap between builders and users
• Anti-values: deciding what AI should not optimise for, and not just what it should do
• Training future AI builders to wear "both hats": cutting-edge engineering plus psychological, environmental and cultural impact
• Co-designing chatbots with communities, with education among the most common use cases in her review of co-design studies
• Her reaction, as a researcher of adult–AI interaction, to parents choosing the sycophantic chatbot in Pilyoung Kim's study (for child-friendly guardrails, see Everyone.AI)
• Should adults form relationships with AI companions? Personalised guardrails that turn empathy down when a user starts becoming over-reliant
• AI value alignment and our "first-order vs. second-order selves": do people know what's good for them?
• Value sliders in chatbot interfaces, convenience culture, and rethinking assessment with AI declaration forms

Pegor Papazian is Chief Development Officer at the TUMO Center for Creative Technologies, the innovative after-school programme for teenagers that was founded in Armenia and has now become a global phenomenon. He trained and practised as an architect before switching to computer science at MIT, where he was a member of the AI Lab.

He founded a Barcelona firm doing computational GIS and urban planning, was a partner in a California software company, and went on to serve as CEO of the National Competitiveness Foundation of Armenia and Head of Program Development at USAID Armenia. He moved to Armenia twenty years ago with his wife, Marie Lou Papazian — TUMO's founding CEO — and their five children. Watching those children enter the Armenian school system became part of the inspiration behind TUMO.

Fifteen years on, TUMO teaches over 35,000 teenagers on a weekly basis across more than 25 centres in 15 countries — including centres in Paris, Berlin, Lisbon, Amsterdam, Los Angeles, Buenos Aires, Takasaki, Astana, Kutaisi, Mumbai, Luanda and Beirut — free of charge, with no grades, no exams, no diplomas and no teachers. In November 2025 TUMO won first place in the WISE Prize for Education for its AI "colearner" platform, which had already taken a top prize in the Learning Engineering Tools Competition. Pegor co-authored "Frankenstein Curricula" with Zachary Pardos for MIT Open Learning's AI + Open Education Initiative, and co-authored TUMO's July 2026 strategy paper, "Making, thinking and learning when capability is no longer scarce."

In this episode, Pegor talks with Svenia Busson about:

• Why capability — writing, coding, illustration — is now commoditised, and why authorship and connection are what became scarce
• "Walk-away pedagogy": the thought experiment behind TUMO — if teenagers could walk away from school without consequence, what would bring them back?
• The AI colearner that deliberately lags behind and asks the student for help — learning by teaching, built into software
• Why TUMO rejected AI tutors, and why the colearner breaks the fourth wall to remind teens it is not a friend
• Why the colearner made students go deeper rather than faster — and what that says about the "two-hour learning" promise
• TUMO's partnership with Anthropic to test the model on maths and reading in Brazil, India and Kenya, and where national curricula may break it
• Why he is against gamification — "Do Not Gamify", his UNESCO Digital Learning Week talk — and the sugar-high problem with streaks and badges
• What 90 teenagers said at ai/teens, the 24-hour conference they ran themselves across 16 cities: AI as a cheat code you use without letting go of the controller
• The licensing-plus-philanthropy model that keeps TUMO free, from Forum des Images in Paris to the State of California in Los Angeles
• The paradox he leaves us with: as AI gets better at teaching something, your reason to learn it from AI disappears

Links mentioned

• TUMO — https://tumo.org
• TUMO's AI work / colearner — https://tumo.ai
• "Making, thinking and learning when capability is no longer scarce" — https://tumo.org/authorship-and-connection
• ai/teens report — https://tumo.ai/teens/report (PDF: https://tumo.ai/assets/pdf/ai_teens.pdf)
• WISE Prize — https://www.wise-qatar.org · Tools Competition — https://tools-competition.org

Zoe Weil is the co-founder and president of the Institute for Humane Education (IHE), which she launched with Rae Sikora in 1996 after teaching environmental and animal-protection courses to twelve- and thirteen-year-olds at the University of Pennsylvania in 1987. She holds an M.T.S. from Harvard Divinity School and an M.A. and B.A. in English Literature from the University of Pennsylvania.

Over nearly four decades, Weil built the field of comprehensive humane education — teaching across the intersections of human rights, environmental sustainability, and animal protection — and created IHE's graduate programs, delivered with Antioch University. She developed the Solutionary Framework, a four-phase process (identify, investigate, innovate, implement) now taught by more than a thousand educators across the US and over 80 countries. She is the author of eight books, including The Solutionary Way (2024, with a foreword by Jane Goodall) and The World Becomes What We Teach: Educating a Generation of Solutionaries, has delivered six TEDx talks, and hosts the podcast Solutionary Voices. Her honors include the NCSS Spirit of America Award (2023) and induction into the Animal Rights Hall of Fame (2010).

In this episode, Zoe talks with Svenia Busson about:

• The Solutionary Framework — a four-phase method (identify, investigate, innovate, implement) for turning students into changemakers
• Why a "solutionary" isn't the same as a problem-solver — and the ethical test that makes the difference
• Parallels with the Design for Change movement and its "I Can" model for youth-led change
• How AI can help students map root causes and stakeholders — without replacing the thinking itself
• Why compassion, built through real contact, is a starting point solutionary work can't outsource to a chatbot
• AI as a "potential nemesis" — manipulation, algorithmic drift, and the slide from healthy skepticism into cynicism
• Turning her own framework on AI itself — who's harmed and who benefits from the technology
• Concrete student wins, from cafeteria reform to rewriting school disciplinary policy