This week's roundup is dominated by evidence and accountability: fresh PISA data on classroom distraction, and a growing chorus of researchers asking schools to slow down and look at what's actually happening in real classrooms. If you only read one thing, make it the piece on how thin the research base still is — handy the next time a vendor tells you "the research says."
New PISA results show 40% of Australian 15-year-olds report being distracted by digital devices during lessons, well above the 28% OECD average, alongside high rates of classroom noise and disorder. OECD education director Andreas Schleicher pointed to AI chatbots and screen time as factors "shrinking attention spans."
Why it matters: It's fresh, credible evidence that unstructured AI and device access correlates with disengagement — useful backing if you're revisiting classroom AI-use routines this term.
A global review of 271 studies across more than 45 countries found only 4 were conducted in Australia, leaving local schools making GenAI decisions largely on overseas evidence. The researchers found outcomes depend far more on task design and teacher guidance than on which tool is used.
Why it matters: Most "AI works" or "AI doesn't work" claims you'll see aren't Australian-tested — a reminder to put PD time into how tasks are designed and scaffolded, not just which tool gets chosen.
The AFT, UFT and Microsoft announced a legally enforceable "National AI Safety & Privacy Standard," barring Microsoft from using student data to train AI models or for advertising, and giving districts control over data retention. It takes effect 1 November 2026, and the AFT says it's now in talks with OpenAI and Anthropic for similar deals.
Why it matters: It's a template worth watching — if similar standards land with other vendors, it changes what a reasonable "AI data privacy" clause should look like in any school's vendor contract.
New York City blocked AI tools for its roughly 600,000 Pre-K–8 students and restricted high schoolers to specific uses; hours later, Los Angeles Unified announced a districtwide moratorium on AI use on school devices for 2026–27, both citing parent and educator concern over shortcuts, weaker critical thinking, and privacy.
Why it matters: When the two largest US districts hit the brakes rather than accelerate, "go slow and get the guardrails right" looks like a mainstream, defensible position — not a lagging one.
Despite huge institutional investment in AI tools, MIT's Justin Reich says large randomised controlled trials are still absent, some peer-reviewed studies claiming AI effectiveness relied on flawed methods, and the models themselves change faster than research can be designed and run.
Why it matters: A handy reality check for the next staff meeting where a vendor cites "the research" — worth asking which research, on which model, tested how.
The 74 Million / Educational Freedom Institute · 17 Sep 2026
A new wave of research is shifting from promotional claims and lab studies to direct, in-classroom observation of how students actually use AI tools day to day, aiming to build an evidence base that can inform curriculum and procurement decisions.
Why it matters: This kind of ground-truth observation should carry more weight than any single vendor case study when you're deciding what to trial next term.
The 2025 PISA results show top performers in China, Japan, South Korea and Singapore pulling further ahead of the US, where reading scores fell to near 25-year lows. Notably, fewer than 10% of students in those top-performing systems reported digital distraction during science lessons.
Why it matters: A useful talking point on tech policy — the highest-performing systems aren't the ones with the most device access, they're the ones with the most disciplined use of it.
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.
