AI in Education Podcast

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

Clinical psychologist and author Andrew Fuller joins Ray and Dan to explore how AI could support neurodivergent learners, strengthen student agency and change the role of teachers.

Andrew argues that neurodiversity should be understood through strengths rather than deficits, and shares practical ways he is already using AI with young people. These include using AI to ask reflective questions, build personalised learning plans, structure revision, prepare for social situations and act as a less confronting source of feedback.

The conversation also looks at the wider implications for schools. If AI can provide more personalised support, what should teachers spend more time doing? Andrew suggests the opportunity is to reduce routine work and give teachers more space to design strong learning experiences, build relationships and develop judgement and curiosity.

They also discuss equity, assessment, AI literacy, school change and why educators may get further by treating AI as something to experiment with, rather than another task to master.

More info

Connect with Andrew on LinkedIn at https://www.linkedin.com/in/andrew-fuller-2238a325/

Andrew's MyLearningStrengths website is at https://www.mylearningstrengths.com/

More info on Andrew's professional services centred on Resilience, psychology, brains and learning for schools, parents, students and professionals at https://andrewfuller.com.au/

And finally, here's where Andrew shares his brilliant free resources for schools, students and parents: https://andrewfuller.com.au/free-resources/

This week on the AI in Education Podcast, Ray is taking a well-earned break, hopefully enjoying a piña colada somewhere sunny while still keeping one eye on the week's AI headlines. Joining Dan as special co-host is Chris Goodall, education leader from the Bourne Education Trust in the UK. Together they unpack one of the most extraordinary weeks of AI news in recent memory, from warnings of existential risk and calls to slow AI development, through to debates about trust, regulation and whether "move fast and break things" is finally coming to an end.

The conversation then shifts to back to education, where the release of the latest PISA results has reignited debates about screen time, technology and student outcomes. Are laptops really to blame for declining results? Or are we falling into the trap of confusing correlation with causation?

Note: This episode was recorded before the news of the OpenAI Australian government Medicare hack, details of which can be found here: Why Australia chose the world's biggest political stage to reveal OpenAI hack

Sources mentioned

• Chris Goodall | LinkedIn
• Screentime limits in public schools to lift our game in the classroom
• Mustafa Suleyman: A Code of Conduct for Humanist AI
• Anthropic CEO Dario Amodei Calls for AI Development to Slow Down
• David McCandless: Information is Beautiful
• Ex-Anthropic Insider Discusses AI Risk on CNN
• Vibe coding - Wikipedia
• Ban the Screens, Keep the Regret: A Rant I Had to Get Out of My System
• Occam's Broom in the Classroom - Lauren Sayer
• Michelle Michaels thought piece - an excellent narrative on the nuiances behind the current news of technology in classroom as always

AI is already changing education. So why are we still trying to hold back the tide? In this episode, Dan and Ray are joined by Dr Phil Cummins, educator, author, co-host of the Game Changers podcast, and Professor of Practice in Education and Leadership at the University of Sydney. Phil argues that education needs to stop "pearl-clutching" over AI and start thinking seriously about what it means for assessment, curriculum and the role of teachers. If students are already using AI, how do we design authentic assessment around the knowledge, skills, dispositions, habits and wisdom we actually want them to develop? The conversation ranges from Phil's beginnings teaching Latin to character education, equity, oracy and the danger of educational fads. At its heart is a powerful argument for the enduring importance of teachers: technology should amplify human relationships, not replace them. For school leaders, Phil's message is refreshingly straightforward: treat AI as a game, learn to use the tools, and find a way to win the game of AI.

Links for more information

A School for tomorrow, Phil's education consulting/research organisation

Game Changers podcast Apple Podcasts: https://podcasts.apple.com/au/podcast/game-changers/id1503430745 Spotify: https://open.spotify.com/show/5WoY7gvWZUWt4Iu3gL2atd?si=41f298a05db2415f

Sydney Character Initiative from the University of Sydney Phil's latest book, On Character

And finally, if you're not sure about the reference to "The Solihull Hospital", then you're in for a treat from the BBC's "Yes, Minister" series: Get some patients - Yes, Minister - BBC

Recorded live at the National Education Summit in Melbourne, this episode explores whether AI in education has reached an inflection point. Ray and Dan unpack a series of recent stories showing how quickly AI is becoming part of wider debates about education, employment, policy and trust. They look at industrial negotiations at the University of Sydney, where staff are seeking consultation over how AI is introduced, and compare that with AI-related provisions already included in Adelaide University's enterprise agreement. As part of the discussion they mentioned The University of Sydney's Cogniti AI, and a previous podcast episode with Dan Hart talking about the university's AI Hub.

They also examine headlines claiming an AI chatbot is "teaching" psychology students at Macquarie University - and the university's response that the tool is supplementary and does not replace academic staff. The conversation then turns to New York City's restrictions on generative AI for younger students, alongside plans for AI critical-thinking education in high schools. Underlying it all is a bigger question: as AI becomes mainstream in education, can the conversation remain balanced - and should the podcast hear more from sceptics, critics and those challenging the direction of travel?

Want to contribute to the discussion? Email us at [email protected]

Want to share your love for the podcast? Well, we can't ask for more than a review in your podcast app, which will help other educators find the stories we share!

What does authorship mean when AI can help us turn our ideas into polished academic writing? And how should universities rethink assessment as those tools become part of everyday life? In this episode, we're joined by Kelly Webb-Davies, Lead Education AI Consultant at the University of Oxford's AI Competency Centre, to explore how Oxford is approaching AI across the institution, from providing tools and training staff and students to building AI capability responsibly. Kelly takes us on her journey in teaching and learning, and challenges some fundamental assumptions about language, writing and authorship. Drawing on her background in linguistics and education, she argues that we need to distinguish between a student's intellectual contribution and the language used to express it. Along the way, we discuss AI detection, linguistic equity, cognitive engagement, oral assessment and why universities may need to rethink what they're actually trying to assess.

Further Reading from Kelly:

Kelly's VFWA Assessment Model https://kellywebbdavies.com/vfwa/

Kelly's DETAP framework https://kellywebbdavies.com/detap/

The AI Inherent Risk Scale (AIIRS) Framework, from Kelly, Dr Mark Bassett and Ella Wicks https://aiirs.ai/

Further reading from other sources:

Kelly talked about the ICAP Framework from ASU https://icap.education.asu.edu/

When talking about writing in your voice, Kelly mentioned Kane Murdoch's blog, which is at https://www.guerillawarfare.net/

And Dan mentioned Sharon Oviatt's book, The Design of Future Educational Interfaces https://www.researchgate.net/publication/286180010_The_Design_of_Future_Educational_Interfaces

Who gets to decide how AI is used in education? Government decides what happens to assessment. Google decides whether Gemini appears by default. OpenAI and parents decide controls for teenagers. Universities and unions decide workplace AI arrangements. Regulators decide expectations around consent. And schools are trying to work out what they should decide.

In this episode, Dan and Ray explore that question through a packed week of AI in education news. They look at Australia's response to AI and take-home assessment, the arrival of parental controls in ChatGPT for Teens, and why Ray is uncomfortable with Google enabling Gemini by default for students. The conversation moves from the classroom to the workplace, with Adelaide University putting AI transparency, human oversight and staff consultation into its enterprise agreement, while new ASQA guidance raises important questions about consent and the use of student data in AI systems. They also unpack new Australian eSafety research into children's use of AI assistants and companions, and tackle the increasingly heated debate about data centres, energy and water use. As Ray puts it, "We're building this plane as we fly it." The challenge is making sure education still gets a say in where it's going.

Here's the links to go deeper into our topics:

Watch: Moment robot beats Usain Bolt's 100m record in China

How education departments across the country are cracking down on students using AI to cheat Education Minister Jason Clare says AI must support learning, not replace it Revealed: The NSW universities using AI in lectures and as 'study buddies'

AI labs begin to muscle in on $6tn education market

Datacentres: The New NSW Regulations

ChatGPT for Teens

Google turns on Gemini AI for students using its classroom app

An Australian first: Adelaide University's new enterprise agreement has explicit provisions on the use of AI in the workplace.

Australia's Vocational Education regulator starts to catch up with responsible AI

Talking to machines: Children's experiences with AI assistants and companions As discussed, good stuff on this from Tim kitchen https://www.linkedin.com/pulse/when-ai-becomes-someone-talk-tim-kitchen-3vznc

What does the research actually tell us about when generative AI helps learning, and when it starts to replace it?

Ray and Dan are joined by Dr Natasha Arthars and Dr Hongzhi (Veronica) Yang from the University of Sydney, and Dr Matthew Hill from the Barker Institute, to explore the new "Young People, Learning and Generative AI" report.

Drawing together a huge body of research, the report moves the conversation beyond whether AI is simply "good" or "bad" for education. Instead, it identifies the conditions that determine whether AI supports genuine learning.

The conversation explores the learner at the centre; the design of the AI tool; the learning task; who else is involved; and the wider cultural and institutional context. Ray learns the difference between cognition and metacognition, and we discuss strategic cognitive offloading, and a crucial new skill for students: deciding what to delegate to AI and what thinking they need to retain themselves.

Ultimately, the question for educators becomes: is AI supporting the learning goal, or replacing the learning?

The Young People, Learning and Generative AI report is online at: https://doi.org/10.25910/89n4-a383 For more project information, and webinar recordings, head to: https://bit.ly/GenAILiterature-Launch

This week, Ray and Dan end up discussing a fundamental question: what is education actually for in the age of AI?

They explore new EU AI Act transparency requirements and the emerging role of AI watermarking, including what this could mean for the often-unreliable world of AI detection.

The conversation then turns to assessment. With Australian policymakers questioning take-home assessment and universities rethinking assessment for an AI-enabled world, Ray and Dan discuss why the goal isn't simply to make assessment "AI-proof", but to ensure students are still developing knowledge, judgement, creativity and critical thinking.

They also examine major new research into young people's use of generative AI from the Barker Institute and University of Sydney. Across 271 peer-reviewed studies, the evidence suggests AI can improve students' immediate performance and engagement - but better performance doesn't necessarily mean better learning.

Finally, they explore "hybrid reading": how students are using AI to orient themselves, clarify difficult concepts, translate academic language and check their own understanding.

Mentioned in this episode:

Matthew Wemyss' guide to EU AI Act transparency requirements Information on Claude's watermarking of AI text - and a geeky read about how it does it New education plugins in ChatGPT Edu Jason Lodge on AI and assessment NSW's news on 'take home' assessments South Australia's Royal Commission into AI Kelly Ilich on teaching standards in the age of AI Danny Liu & Adam Bridgeman's "What Should Education Look Like in the Age of AI?" Barker Institute's research paper "Young People, Learning and Generative AI" Research from Deakin University's CRADLE team on hybrid reading practices

The Hidden Gender Gap in AI Education, with Juliana Peloche In this episode of the AI in Education Podcast, we welcome Juliana Peloche to explore one of the most important (and often overlooked) conversations in educational AI: the gender gap. Juliana shares how a Year 5 student's curiosity about AI sparked her own research journey, leading to a PhD and, ultimately, the co-founding of Women in AI Australia. Together, we discuss why representation matters (see Will Artificial Intelligence Get in the Way of Achieving Gender Equality?), the barriers women continue to face in AI (see HBR's Women Are Avoiding AI. Will Their Careers Suffer? and Global Evidence on Gender Gaps and Generative AI Over Time) and what educators, institutions and industry can do to create more inclusive opportunities. The conversation hears about new research into how female university students are using AI, revealing surprising findings around ethics, trust, academic integrity, and confidence with AI tools. We also explore AI bias, the importance of visible role models, and practical ways schools and universities can encourage more women to lead in the field. Links and resources: Women in AI Australia Learn more about the Women in AI Australia upcoming events Connect with Juliana Peloche on LinkedIn Go deeper: Ray's free AI for Humans course covers this in "Bias in, bias out - spotting unfair patterns"