This week's reading has a clear thread: the evidence is catching up with the enthusiasm, and it's asking us to slow down and think about trust, oversight and what learning actually requires. From a large AI tutor trial with sobering results to new Australian data on how little parents check their kids' AI use, there's plenty here to bring to your next staff conversation.
A TrendLife survey of 1,030 Australian parents found 75% say their children use general AI tools for schoolwork, including 32% reporting regular use among 5–12 year olds, yet only 42% regularly review or discuss that use. Privacy and security (58%), accuracy (47%) and over-reliance (45%) were parents' top concerns.
Why it matters: If a third of primary-aged kids are already using AI at home with little oversight, schools have a real opening to run parent AI-literacy sessions and send home simple 'talk about it' prompts.
Responding to NSW's new public-school screentime limits (15 minutes a day for K–2, rising to 45 minutes for Years 5–6, from 2027), Twinkl's James Lister backs the caps but argues the better question is why and when a screen is used. He also warns that enthusiasm for generative AI shouldn't outrun the evidence, especially where AI removes the productive struggle learning depends on.
Why it matters: With K–2 screen minutes about to become scarce in NSW, every digital or AI activity will need to clearly beat the pen-and-paper alternative, so now is a good time to audit which tools actually earn their place.
CQUniversity's Dr Ragnar Purje argues that critical thinking and higher-order reasoning take years of human interaction to develop, and that there are 'no shortcuts' in brain development. He points to OECD findings that generative AI can lift task performance without producing genuine learning gains, and calls for schools to protect face-to-face interaction and physical play.
Why it matters: It's a useful counterweight for planning conversations: for younger learners especially, AI use should be weighed against what it displaces (talk, play, effortful thinking), not just what it speeds up.
Annenberg EdWorkingPapers (Univ. of Maryland) · 30 Sep 2026
In a randomised trial with 2,379 undergraduates and 30 instructors, students given a course-integrated generative AI tutor ended up with final grades 0.37 standard deviations lower and much lower participation in existing course activities. First-generation students saw the largest losses.
Why it matters: It's a university study, but the lesson carries to K–12: an AI tutor can quietly displace the practice and participation that actually build learning, so any rollout needs a plan for how it fits alongside existing routines, not instead of them.
A Center for Digital Thriving report found 74% of teachers and 69% of principals framed their AI dilemmas around cheating, while interviewed teens described fearing false accusations, deliberately underperforming to avoid suspicion, and bypassing teachers for chatbots. Researchers recommend building AI policies around trust, with students involved in shaping them.
Why it matters: Swapping 'Did you use AI?' for open, curious questions about process, and co-designing class AI norms with students, protects the relationships that make learning (and honesty) possible.
Erin Mote, CEO of InnovateEdu, who once banned social media at the school she co-founded, argues against outright AI bans and instead for targeted rules. Her list: ban AI companion chatbots, plan responses to deepfake images, set clear assignment-level AI rules, keep first drafts human, and require independent evaluation of AI products.
Why it matters: It's a concrete, ready-made checklist for any school writing or reviewing its AI policy, and it separates the genuinely harmful uses from the useful ones.
A survey of 1,405 K–12 teachers in the US, India, Qatar, Colombia and the Philippines found AI readiness, school support and prior use explained half the variation in positive beliefs about AI, but almost none of the variation in plagiarism and creativity concerns. More AI-competent teachers actually reported higher creativity concerns.
Why it matters: PD that builds confidence won't automatically ease teachers' concerns, so training should make room for the ethical and creative-thinking questions experienced users raise, rather than treating them as resistance.
A five-minute read for busy teachers: the week's most useful AI-in-education news, Australian stories first, each with a note on why it matters in your classroom. Free, and you can unsubscribe at any time.
