AI for every learner: lessons from learners with additional learning needs

Libby Hills is co-founder and CEO of EdTechnical. It started as a podcast and now gives revenue-based financing and research support to AI edtech companies with $1–5M in annual recurring revenue (ARR). She holds an MSc in Education (emerging technologies and AI) from the University of Oxford. She started her career as a Teach First classroom teacher in London, after turning down an approach from a British intelligence agency.She went on to lead technology pilots across PEAS, a network of low-cost schools in Africa. Then, at the Jacobs Foundation, she co-led the global edtech investment portfolio and built an evidence-assessment model based on the US ESSA framework. Her article “Evidence pays” argues that proof of impact is a commercial imperative and not just an academic exercise. In two Jacobs funds, companies with evidence-backed products saw 50–70% higher returns. Since 2023 she has co-hosted the EdTechnical podcast with Owen Henkel, with more than 57 research-grounded episodes on AI and learning.In this episode, Libby talks with Svenia Busson about:Going from early adopter to sceptic to realist, a path that follows Gartner’s Hype Cycle. Plus the question from Wendy Kopp that changed how she saw classroom tech: “How do you know kids are really learning?”Evidence as a journey, not one high-stakes RCT: logic models for a few thousand dollars, AI output-quality “research sprints”, LeanLab Education’s rapid-cycle testing and WhatWorked Education’s mini-RCTsWhy evidence pays: buyers, regulators and the edtech backlash are making proof of impact a competitive advantageWhat we actually know so far: teachers save time (the EEF’s RCT of Oak National Academy’s Aila), but learning gains depend on design. Generic chatbots remove the productive friction that learning needs.Pushing back on the Brookings “pre-mortem”: unsupervised chatbot use is not the same as purpose-built education toolsAI toys and social robots: clunky today, but the next generation could create parasocial bonds with very young childrenAlpha School, microschools and why public schools won’t look very different in 20 years. Procurement, contracts and timetables act as a brake on change.The end of graded take-home essays. Professor Panos Ipeirotis at NYU Stern built AI-powered oral exams. Plus three questions every school leader should ask an AI vendor.

Jöran Muuß-Merholz is founder and co-managing director of J&K – Jöran und Konsorten, an education agency that calls itself a “think-and-do tank for contemporary learning.” He trained as an educational scientist (Diplom-Pädagoge) and learned media education as an apprentice to Professor Stefan Aufenanger.He set up the Friedrich Naumann Foundation’s education office in Hamburg. He then co-founded the Archiv der Zukunft network with filmmaker Reinhard Kahl and ran it as managing director. It grew to more than 1,000 members, and its 2008 congress “Treibhäuser & Co” in Bregenz drew 1,400 participants. In 2009 he founded J&K. In 2012 he co-founded OERcamp, Germany’s first barcamp on Open Educational Resources. He translated the OECD Learning Compass 2030 and Fadel, Bialik and Trilling’s Four-Dimensional Education into German. In 2026 he also translated the German summary of the UNESCO Global Education Monitoring Report.He writes a column for the German journal PÄDAGOGIK and has spoken in about 20 countries. His latest book is Schule 2035: Lernen nach Digitalisierung & KI (Beltz, 2025).In this episode, Jöran talks with Svenia Busson about:Why imagine schools exactly ten years ahead: close enough to act on, far enough to free the imagination. The idea came from Florence Gaub’s Zukunft. Eine Bedienungsanleitung — and it’s also why “AI is just a tool” is not enough.LIDA (“Learning Individually, Digital, Adaptive”), a school where every student works alone with an AI tutor — and the unsettling finding that students at Minerva University felt comfortable being monitored all the timeWhat if some teenagers simply prefer learning alone with a machine? And what does it say about adults if we overrule them?The Manfred Spitzer Lyceum: digital-free and back to basics, named after the author of Digitale Demenz (“Digital Dementia”), with a bakery, a printing press and a large gardenThe Unschule (“anti-school”): built to protect the joy of learning, as a response to the Lee Jenkins curve, inside a converted shopping mall where AI closes the gap between a child’s question and the learning — even after 7 of its first 11 graduates fail their final examsPallas Athene: a network of 50 schools sharing 1,000 online experts, “the Minerva idea for 14-year-olds,” with AI scheduling 17,000 studentsThe Lisa Rosa Reform School: projects built around the SDGs (the UN Sustainable Development Goals) with partner schools on other continents, technology used to handle complexity rather than reduce it — and a shared calendar instead of fixed classroomsFutures thinking as the blind spot of school development: only 0–1% of pages in school-development books are about vision. Also discussed: UNESCO Futures Literacy and Rachael Smith’s work at The Education FuturistFive imagined schools, no villains: a starting point for picturing your own school of the future.

Bukky Yusuf FCCT FRSA is a deputy headteacher and science lead at a small, deliberately low-sensory special school in London for 14 to 19 year olds. She is also a leadership development coach and education consultant. She trained as a secondary chemistry teacher through a school-based PGCE after her master’s. In 2019 the Department for Education appointed her co-chair of its EdTech Leadership Group to support the government’s edtech strategy.She has spent more than 20 years in education. She taught in mainstream schools across London, worked as a secondary science consultant for a local authority, and held senior leadership roles leading professional development, quality-first teaching and whole-school technology rollout. She moved into special education to close the gap she kept seeing between students with additional learning needs and their peers. She planned to stay two years and is still there. Today she reviews and assesses schools for the Good Future Foundation’s AI Quality Mark. She sits on Innovate My School’s Education Board, judges the EdTech50 Schools and BETT Awards, and is a WomenEd network leader and Leadership Matters ambassador. She co-wrote The Big Book of Whole School Wellbeing (Corwin, 2021) and wrote a chapter of The Early Career Framework Handbook (Chartered College of Teaching, 2020).In this episode, Bukky talks with Svenia Busson about:The quiet advocates behind a career: a PGCE mentor, a headteacher who looked past an imperfect interview, and the coaches who followedWhy more students have SEND and complex needs: the post-lockdown “unmasking”, rising mental health needs, and devices built to keep us doomscrolling 24/7Inside a 20-student, low-sensory special school, where young people who didn’t fit mainstream go on to GCSEs, A levels, university and apprenticeships — and why England’s SEND reform won’t go far enough without more teachers, teaching assistants and better trainingStudents who reject AI: worries about environmental justice, disdain for “AI slop”, and seeing teachers’ AI use as cheatingHow the Good Future Foundation’s AI Quality Mark works: nine areas scored 1 to 5, self-assessment checked against evidence, and awards from Bronze to Gold — plus what the AI Shift in Schools report says about adopting AI responsiblyA sensible first year for schools: start with the tools you already have, go for workload quick wins, keep a policy you review every term, use Dr Leon Furze’s AI Assessment Scale, and give students with additional needs a voiceEvaluating edtech before you buy, with the Chartered College’s evidence questions — and why “free” still costs time you can’t get back“The notes between the notes”: José Picardo’s argument that AI may do for teaching what photography did for painting, pushing us toward what is most humanA grounded, equity-first voice on AI for every learner — especially those mainstream misses.Resources mentionedDr Leon Furze, Updating the AI Assessment Scale: https://leonfurze.com/2024/08/28/updating-the-ai-assessment-scale/The AI Shift in Schools: Risks, Realities and What’s Coming Next (Goodnotes research publication): https://cdn.prod.website-files.com/69c104545fc04cbc4f88ac57/69da617bb984fb0dc4cfa8ab_68ee2209046a4f0ad751be24_The%20AI%20Shift%20in%20Schools.pdfJosé Picardo, The Notes Between the Notes: What AI Is Actually For (Azimuth): https://www.azimuth.org.uk/the-notes-between-the-notes-what-ai-is-actually-for/Chartered College of Teaching, EdTech Check and evaluation resources: https://chartered.college/edtech-evidence-board-project/resources/Keeping Children Safe in Education (KCSIE): https://www.gov.uk/government/publications/keeping-children-safe-in-education--2Previous episode with Daniel Emmerson (Good Future Foundation), Ep 83

Daniel Emmerson is the founding Executive Director of the Good Future Foundation, a UK-based charity helping schools around the world adopt AI strategically and responsibly. He is also Academic Affairs Lead at Goodnotes. Before moving into AI and education, he spent 14 years in schools as a teacher and senior leader. Most notably, he was Director of Global Education at Felsted School, where he built cross-cultural learning programmes for more than 4,000 students across 40 countries.In 2023 he led Goodnotes’ research project “Generative AI in Schools: Evolving Coursework, Homework and Assessment in the Age of AI”, built on conversations with school leaders, teachers, students and support staff worldwide. Its findings led him and Goodnotes founder Steven Chan to launch the Good Future Foundation in April 2024. The charity is fully tech-agnostic and is governed by an advisory council and a student council (aged 15 to 22) with equal oversight. Its flagship AI Quality Mark was co-designed with school leaders including Dr Andy Kemp of the National Mathematics and Science College. It now supports more than 6,000 schools, with partners in South Africa, Ghana, Argentina and South Korea.Daniel also hosts the Foundation’s podcast, Foundational Impact, sits on the board of the Institute for Global Learning, and serves as a school governor in Essex.In this episode, Daniel talks with Svenia Busson about:From a 2023 Goodnotes research project to a charity — and why it never recommends a single AI toolInformed refusal vs. burying your head in the sand: why ignoring AI is the real risk for schoolsThe AI Quality Mark’s nine pillars, 1–5 self-assessment and Progress, Bronze, Silver and Gold awards — rewarding the thinking process, not the amount of AI usedHow a Gold school can choose not to use AI for marking after listening to its students — and still score highlyThe “efficiency trap”: why AI rarely saves teachers time — and why starting with purpose matters moreBeyond AI-detection “whack-a-mole”: marking prompts, AI-use declarations and teachers modelling their own AI useA student council’s warning about losing productive friction and the zone of proximal developmentDurable skills for an AI world: perspective-taking, cross-cultural collaboration — and why “a carpenter with AI is also an architect”Links mentionedGood Future Foundation — https://www.goodfuture.foundationAI Quality Mark — https://www.goodfuture.foundation/ai-markFoundational Impact podcast — https://www.goodfuture.foundation/podcast“Generative AI in Schools: Evolving Coursework, Homework and Assessment in the Age of AI” (Goodnotes, 2023) — https://www.goodnotes.com/research/generative-ai-in-schoolsInstitute for Global Learning (formerly the Global Education Benchmark Group) — https://gebg.orgThe Big AI Project (Big Education) — https://bigeducation.org/the-big-ai-project/

Louis de Diesbach is a technology ethicist, author and PhD candidate in Business & Technology Ethics at the Solvay Brussels School of Economics & Management (ULB). He studied Latin and Greek in high school, then management sciences at Solvay, and later went back to earn a Master’s in Ethics and Philosophy at ULB. A mentor, the corporate philosopher Luc de Brabandere, showed him that philosophy and business could live in the same world.After about ten years in consulting, including at Boston Consulting Group, Louis now works on AI ethics and governance in the corporate world. He co-founded the ULB Research Lab in Business & Technology Ethics, lectures on technology ethics at HEC Liège, and sits on the ethics committee of the computer-vision company XXII. He has written three books: Liker sa servitude (FYP, 2023), Bonjour ChatGPT (Mardaga, 2024, with a foreword by Luc de Brabandere) and Faussaires algorithmiques: l’IA va-t-elle remplacer les artistes ? (Éditions de l’Aube, 2026), which draws on more than 120 conversations with artists, philosophers and lawyers. He writes columns for L’Écho and La Libre Belgique and gave the TEDxNantes talk “Osons penser une nouvelle utopie numérique”.In this episode, Louis talks with Svenia Busson about:Why tech companies need ethicists, and the “ethics washing” behind 1,000+ job openings at OpenAI, Anthropic and Google, none of which required a philosophy degreeSuperintelligence talk as a distraction from AI’s real costs: energy use, scraped creative work, and the hidden labour of data annotatorsThe cognitive-debt paradox: executives who discuss AI’s risks in the morning, then have ChatGPT do the afternoon exerciseCamus’s 1957 letter to his teacher Louis Germain, and why replacing teachers with AI misunderstands what teaching isWhere AI does help (dyslexia and dyscalculia support, mixed-level classrooms, drafting test questions) and the case for small, specialised models over all-purpose LLMsThe Max Planck study showing people now say “delve” more often, and what the homogenisation of language means for students still finding their voiceSix lenses school leaders should apply before adopting AI, with training as the foundation, and the three Cs students need: critical thinking, collaboration and creativityScreens, young kids and AI companions, from Her and Black Mirror to Simone de Beauvoir’s question: what will we do with all the time we save?

Dana Suskind is Professor of Surgery and Pediatrics at the University of Chicago and a pediatric cochlear implant surgeon. She earned her BS and MD at the University of Missouri–Kansas City and joined UChicago in 2002, where she founded the Pediatric Hearing Loss and Cochlear Implant Program.Her patients sent her in a new direction. Two children could get identical implants and still have wildly different outcomes, and she saw that the difference was the language around each child, not the device. She left the operating room part-time and founded the Thirty Million Words Initiative in 2010. In 2017 she co-founded the TMW Center for Early Learning + Public Health with economist John List. The center supports parents and caregivers with programs on the power of talk and interaction, and argues for the policies families need. She wrote Thirty Million Words (2015) and Parent Nation (2022), an instant New York Times bestseller and Nautilus Gold winner. Her new book, Human Raised: Nurturing Connection, Curiosity & Lifelong Learning in the Age of AI (Dutton, 2026), asks what it takes to raise children well now that machines can mimic the interactions that build their brains.In this episode, Dana talks with Svenia Busson about:Two children, identical cochlear implants, opposite outcomes — and what that taught a surgeon about early language environmentsThe “helpless infant paradox” and brain-to-brain synchrony — why human connection is biological, not just a verbal back-and-forth (drawing on Ruth Feldman’s research)“Good enough” parenting — how friction, missteps, rupture and repair build empathy, theory of mind and frustration toleranceThe HOPE framework: Human connection, Own your imperfections, Protect the early years (and learning), Enhance not replaceSycophantic, frictionless AI and children who learn that relationships revolve around them — “no bots for tots,” versus science-led socially assistive robots and robotic reading companionsThe DETECT method for judging any AI product for children — and a “Developmentally Aligned” seal modelled on the Pure Food and Drug Act and the Good Housekeeping SealThe inequality trap — human connection as “the new organic,” and why a low-tech childhood is the best preparation for a high-tech futureRethinking what schools are for — human-edge skills, and investing in the parks, museums and civic spaces that bring people togetherA surgeon’s case for keeping childhood human while the machines get smarter.Resources mentionedHuman Raised, Dana Suskind (Dutton, 2026)Parent Nation, Dana Suskind with Lydia DenworthTMW Center for Early Learning + Public Health, University of ChicagoAmerican Academy of Pediatrics, “Digital Ecosystems, Children, and Adolescents” (2026)Common Sense Media, AI Toys Risk Assessment (Jan 2026)Sarah Sebo’s lab (UChicago), robotic reading companions and reading anxietyBrian Scassellati (Yale), social robots and autismMaja Matarić (USC), socially assistive robotics

Dr. Sven Schütt is CEO of IU Group, home to IU International University of Applied Sciences, the largest university in Germany. He trained as a neuroscientist, studying physics and philosophy in parallel and finishing both in the standard time. He then earned a PhD on learning at the Max Planck Institute of Neurobiology during the “AI winter.”After McKinsey, he became Managing Director at Germany’s Federal Employment Agency, where he helped implement the Hartz labour-market reforms, one of Europe’s largest public-sector reforms. He then ran the Hertie School in Berlin. Since 2010 he has turned IU from a 1,500-student hotel-management school into a group with about 130,000 students in Germany and 20,000 more worldwide, with institutions in the UK, Canada and Austria. More than 70% of IU students come from non-academic households, and its dual-study model has 15,000 corporate partners. IU was the first university in the world to give every course an AI companion built on GPT-4. That tool has since grown into Syntea, a multi-agent learning companion.In this episode, Sven talks with Svenia Busson about:Building a university around working adults: start any day, sit an exam any hour — and why access is the core missionInside Syntea: from a Q&A chatbot to a companion that maps your goals, books your exams and nudges you back on WhatsAppWhat IU’s data shows: 58% usage in a 77,000-student study, about 25% fewer early dropouts, about 30% faster learning — and how close that is to Bloom’s “two sigma” effectIf AI cuts learning time by a third, should a three-year bachelor’s shrink to 18 months? — and why regulation hasn’t caught up“From grading to guiding” and cognitive sovereignty — why universities teach judgment, not knowledge (Samuel Johnson declared them obsolete in 1766)The AI gender gap: men use AI about 30% more than women — and how organisational encouragement nearly closes it, as AI fluency enters more than 3,000 coursesThe crisis of assessment: proctored colloquia, personalised AI-driven exams and AI-supported grading for fairnessThe future of teaching: the end of passive lectures, “augmented intelligence,” and why research may be disrupted more than teachingA university CEO arguing that every institution must now become a tech company.Links & resources mentionedIU International University of Applied Sciences: https://www.iu.orgSyntea, IU’s AI learning companion: https://www.edtechinnovationhub.com/news/iu-expands-syntea-into-ai-learning-companion-for-80000-distance-studentsIU’s 77,000-student analysis of AI learning-assistant use (AIET 2026): https://www.iu.de/news/en/worlds-largest-analysis-of-ai-use-in-higher-education-shows-ai-supported-learning-has-become-part-of-everyday-student-life/University Experience Study 2026 (AI fluency of IU students): https://www.iu.de/news/en/study-on-ai-in-higher-education-students-learn-faster-and-more-efficiently/IU Job Readiness Framework: https://www.fernstudi.net/news/iu-hochschule-startet-flaechendeckendes-ki-kompetenzmodell-in-der-lehreIU AI Research Institute (Prof. Thorsten Fröhlich, Prof. Kamal Bhattacharya): https://www.iu.de/news/en/iu-international-university-of-applied-sciences-establishes-new-ai-research-institute/Benjamin Bloom, “The 2 Sigma Problem” (1984): https://journals.sagepub.com/doi/10.3102/0013189X013006004Sven Schütt on LinkedIn: https://www.linkedin.com/in/sven-sch%C3%BCtt-04a24140/

Isabelle C. Hau is the Executive Director of the Stanford Accelerator for Learning. It’s a Stanford-wide effort hosted at the Graduate School of Education, and it brings together more than 100 faculty affiliates to connect learning science with practice and policy. She grew up in rural southern France and studied at Sciences Po Paris and ESSEC before earning her MBA at Harvard Business School. She began her career as a financier at Morgan Stanley.Before Stanford, she spent about a decade with the Omidyar Group. She founded and led the US Education initiative at Omidyar Network and was a founding partner of its spinout, Imaginable Futures, where she backed early-stage organizations including Khan Academy Kids, Ellevation and Handshake. She wrote Love to Learn: The Transformative Power of Care and Connection in Early Education (PublicAffairs, 2025) and the Stanford Social Innovation Review cover story “Welcome to the Era of Relational Intelligence.” She also serves on the Brookings Institution’s Global AI Taskforce and the EDSAFE AI Alliance steering committee, and she writes the weekly newsletter Small Talks.In this episode, Isabelle talks with Svenia Busson about:Relational intelligence (RQ) as the next stage of human intelligence after IQ and EQ, and why we need to decide on purpose what our humanity should becomeThe “relational recession”: fewer siblings, fewer friends, less play and less contact between generations, and how AI companionship is moving into that gapAI sycophancy: new Stanford and Oxford research on how chatbots that constantly agree with us make real human relationships feel harder and less rewardingAn 8-workshop pilot at Design Tech High on how to build relationships, including two sessions on conflict and repairAI literacy through tinkering instead of a fixed curriculum: the Stanford AI Tinkery and its new free network of schoolsEducation’s “Turing Trap” and what the latest PISA results show: AI without scaffolding hurts outcomes, while AI with good teaching support helpsWhere AI is showing real promise: comic-book science projects that made students more curious, an autism screening game that is 86% accurate, and Tutor CoPilot, which brings novice tutors up to the level of the best onesWhy AI-only tutors struggle to get used, how to measure relationships without reducing them to a score, and what K-12 can learn from medical educationResources mentionedIsabelle’s writingHau, I. C. (2025). Love to Learn: The Transformative Power of Care and Connection in Early Education. PublicAffairs. LinkHau, I. C. (2026). “Welcome to the Era of Relational Intelligence.” Stanford Social Innovation Review, 24(2), 26–35. LinkHau, I. C. (2026). “Why Human Connection Must Be at the Heart of Education in the AI Age.” World Economic Forum. LinkHau, I. C., & Schwartz, D. L. (2025/2026). “Avoiding Education’s Turing Trap.” Stanford Social Innovation Review. LinkHau, I. C. (2026). “Curiosity, Connection, and the Future of Learning.” Educational Leadership, 83(7). LinkResearch on AI sycophancy and relationshipsCheng, M., et al. (2026). “Sycophantic AI decreases prosocial intentions and promotes dependence.” Science, 391(6792). LinkIbrahim, L., et al. (2026). “Sycophantic AI makes human interaction feel more effortful and less satisfying over time.” arXiv:2605.07912. LinkKim, P., Xie, Y., & Yang, S. (2025). “‘I am here for you’: How relational conversational AI appeals to adolescents, especially those who are socially and emotionally vulnerable.” arXiv:2512.15117. LinkResearch on AI tutoringTurano, C., Pihl, V., Agnew, C., Ziegler, L., & Loeb, S. (2026). AI Tutoring Is Not a Monolith: What We Actually Know. SCALE Initiative, Stanford Accelerator for Learning. LinkTutor CoPilot, Stanford National Student Support Accelerator. LinkStanford initiativesStanford Accelerator for Learning. LinkStanford AI Tinkery. Link

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 toolsBuilding a coalition of 50+ partners around shared values, and keeping commercial and political motives out of the roomA four-hour, AI-agnostic training for heads — and why their confidence rises from 1–2 out of 5 to 4–5 out of 5AI literacy from age 3 to 14, with prediction games, bias and stereotyping activities, and class agreements not to share misinformationAnthropomorphism, 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 trainingWhy human connection and essential skills matter more than ever in a world full of AIThe 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 apartWhat causes dyslexia: phonological processing, family history, and why reading ability sits on a continuumWhy decades of reading science still don't reach classrooms, with the US as a case study: only ~40% of fourth graders read proficiently on NAEPWhy 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 outrightInside the EVAL Collaborative: pre-registered, open-science RCTs published regardless of outcome, and the first EVAL AI in Learning ChallengeCan 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 timeAdvice 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