AI in Education Podcast

What if giving students access to AI tutors actually makes them learn less?
This week, Ray and Dan examine two new randomised trials that challenge some of the assumptions about AI in education. One found that AI tutoring was as effective as expert human tutoring, but students covered less material and saw no improvement in final test scores. Another found that students offered AI tutors performed worse, participated less, and rarely used the tools.
We explore what these findings mean for student motivation, engagement and the future of teaching.
News this episode:
Universities embrace ChatGPT: Princeton and Carnegie Mellon announce wider access, while Dartmouth's provost resigns amid controversy over AI-written content.
Google changes course: Gemini will no longer be enabled by default in schools, raising questions about AI and Google Search.
Google's blog, and follow up clarification tweet
Reporting on it: NY Times story and The Atlantic Story
Rethinking assessment: Microsoft's medical AI highlights why traditional exams may not reflect real-world skills.
AI policy in Australia: Major changes to national education bodies and renewed debate about screen time and testing.
Research
Measuring the impact of AI tutoring, from near transfer to standardised tests: An expanded RCT in UK mathematics classrooms [Blog post] [Research Paper]
AI humanizers and the political economy of academic integrity in higher education
The Effects of Course-Integrated AI Tutoring on Student Performance and Engagement: A Randomized University Trial

What happens to maths education when AI becomes better at mathematics than the mathematicians themselves?
In this episode of the AI in Education Podcast, Ray and Dan are joined by Rita Bateson, founder of Eblana Learning and an expert in curriculum, assessment and ethical AI, to explore how AI's extraordinary mathematical advances are challenging everything we thought we knew about teaching and learning.
From AI solving complex mathematical problems to the future of STEM careers, Rita explains why maths teachers are suddenly confronting the same disruption that humanities teachers have faced for years.
The conversation also explores:
Rethinking maths education: Why AI demands new approaches to curriculum, assessment and mathematical understanding.
Students rejecting AI slop: Why young people want authentic teacher feedback, not polished AI-generated responses.
AI anxiety and trust: What students really think about AI and why schools need more honest conversations.
Practical opportunities: How teachers can use AI to rediscover ideas, unlock creativity and improve teaching without sacrificing human connection.
Links
Eblana Learning website: eblanalearning.com 
Rita's Substack: https://substack.com/@eblana  
Rit'a book for school leaders:  https://www.hachettelearning.com/teaching-strategies/international-school-leaders-guide-to-ai  
The new workbook by Eblana Learning with Pearson for ethical AI literacy:  https://www.pearson.com/international-schools/international-baccalaureate-curriculum/ib-ai-student-workbooks.html  
After we recorded this podcast, Hannah Fry did an excellent explainer about the AI plus Maths debate in the Rest is Science podcast which came out straight after we spoke: Apple Podcasts and Spotify

Is AI actually improving learning, or are we making things easier without making students better learners?
This week we explore five research studies challenging assumptions about AI in education. From AI-assisted grading and student motivation to the surprising finding that Socratic AI tutors can leave students more frustrated and performing worse, the evidence raises important questions about how we design AI-supported learning.
We also explore:
AI and assessment: Why AI-generated feedback might be more valuable before students submit their work.
Learning independence: Does relying on AI reduce students' willingness to tackle problems alone?
AI tutors: Why asking more questions doesn't necessarily produce better learning outcomes.
Teen safety: Common Sense Media's concerning findings about ChatGPT's protections for teenagers.
AI adoption: Arden University's Gemini rollout and Microsoft's push to support vocational education teachers.
Plus, why researchers need to compare AI grading against human-to-human marking consistency before drawing sweeping conclusions.
 
Links to News and Research
Arden University go with Google Gemini
https://www.prnewswire.com/news-releases/arden-university-and-google-cloud-form-strategic-partnership-to-ensure-development-of-job-ready-ai-native-graduates-302901909.html 
Common Sense Media test ChatGPT's for Teens
https://www.commonsensemedia.org/articles/chatgpt-claims-to-have-a-safety-net-for-teens-we-watched-it-fail 
The Guardian: "As more Australian universities use AI to mark students' work, is it creating a 'slop-cycle'?"
https://www.theguardian.com/australia-news/2026/sep/28/university-ai-marking-academic-overworked-staff-student-shortcut 
AI for VET Educators National Program
https://www.tech-academy.com.au/aiforeducators 
Dan mentioned the webinar he hosted with ASI on building AI agents in Copilot Studio and Foundry.
Here's the link: https://www.youtube.com/watch?v=idVDgx6nH1M 
Research
The GenAI Motivation and Learning Buddy: A Generative AI Prompt Library for Educators and their Students
https://zenodo.org/records/20533907 
Can large language models reproduce higher education grade bands? Cross-model study of calibration and grading bias in authentic student writing
https://www.tandfonline.com/doi/full/10.1080/02602938.2026.2734796 
Generative AI and Learning Dynamics in Higher Education: A Longitudinal Empirical Study
https://onlinelibrary.wiley.com/doi/10.1002/jcal.70322 
AI Assistance Reduces Persistence and Hurts Independent Performance
https://arxiv.org/abs/2604.04721v5 
Guardrails or Roadblocks? Effects of Pedagogical Style and Context Awareness in AI Teaching Assistants for Programming
https://arxiv.org/abs/2609.29995

What if AI could help us build schools that don't just prepare students for the future, but fundamentally rethink what education should look like?
In this episode, Ray and Dan are joined by Will Campbell, Founding Head of Franklin School in New Jersey, a school recognised internationally for its innovative approach to education and winner of 2025 World's Best School Prize for Innovation by T4 Education
Will shares how Franklin has embraced AI from the outset, moving beyond fears of automation to create a learning environment where technology gives teachers time back, challenges students to think more deeply, and develops the human capabilities that matter most.
The conversation explores:
Building the school of the future – rethinking curriculum, timetables and traditional education models.
AI and teacher empowerment – how Franklin aims to save teachers 8–9 hours a week while strengthening their professional judgement.
Rethinking assessment – why AI should make learning more challenging, not easier.
Developing superhuman skills – prioritising critical thinking, ethics, resilience and real-world experiences.
Personalised learning – moving beyond age-based progression towards skills, capabilities and student autonomy.
An inspiring conversation about building schools where AI enhances human potential rather than replacing it.

In this episode Ray and Dan explore a whirlwind of AI announcements and research, asking a critical question: Is AI helping students learn, or simply helping them produce better work?
Australian universities are going all-in on AI, with UNSW, Flinders and the University of Sydney announcing major rollouts. Meanwhile, New South Wales schools are being warned not to rely on AI detection tools to identify cheating.
The research raises fascinating questions. Estonia's nationwide AI experiment reveals surprising student resistance to AI tutoring, while OECD PISA data highlights concerning links between frequent AI use and lower assessment scores. New research also explores the Learning Performance Paradox - when AI improves students' work without improving their understanding.
Ray and Dan also discuss screen time in schools, the disappearance of Custom GPTs and Gems, and why much AI education research needs greater rigour.
Plus, Ray announces that his complete AI for Humans course is now available free!

Links and resources
Uni's go all-in on AI tools - with UNSW, Flinders, The University of Sydney and Western Syndey University all jumping on the AI bus

NESA say "No way" to AI Detectors

UNESCO - The Algorithm in the Room

OECD - PISA research on AI use and student performance

Estonia - AI Leap programme findings

NSW research - Dr Jo Orlando's school screen-time study

Research - In Vino, Veritas, and Vulnerabilities

Research - Stanford's Understanding the Evidence Base on AI in K-12 Education
Research - Generative AI in Legal Education: A Two-Year Experiment

Research - Building AI companions that prioritise learning over performance
And finally, did we mention this enough?
AI for Humans - Free AI fluency course