AI in Schools
Victoria mandates device-free time in every secondary school
Digital Attitudes, 18 June 2026
The Thursday read is one Victorian cabinet announcement that puts device-free time on every secondary school's term 1 2027 timetable, one Melbourne data conference opening today on what schools should actually do with the dashboards they already pay for, and one piece from the academic press on the AI tools the early years sector is already procuring without the guidance the school side has spent two years building.
Carroll's device-free time mandate lands on every Victorian secondary school for term 1 2027
Deputy Premier and Minister for Education Ben Carroll announced on Tuesday 16 June that every Victorian secondary school will be required to include planned device-free time in the curriculum from term 1 2027, with the operating rule that any in-class device use has to be intentional rather than default. The policy sits on top of the existing primary cap (90 minutes per day for years 3 to 6, minimal time for prep to year 2) and the 28 January 2027 device ban covering phones, smart watches, and wireless earbuds across every Victorian school, government and non-government. Carroll's framing: "We know that screen-free time is important to help students engage with their learning, so we're asking schools to include it in their curriculum." Inside State Government on the secondary school device-free time announcement, 16 June.
Why this matters: For Victorian independent and Catholic school heads briefing the council's IT investment paper against the term 1 2027 commencement, the operational read is that the BYOD program (the laptop class set parents currently buy) now sits in the same compliance bucket as the personal phone policy, and the question the term 4 council meeting answers is what the actual in-class device use is once "intentional" replaces "default." For state Department of Education CIOs running the digital learning stack across the 1,500 government secondary schools, the operational read is that the LMS minutes the existing Compass, Sentral, and Schoolbox installs already log against each lesson are now the audit trail the 2027 compliance frame will be read against, and the procurement question the October budget paper answers is whether the existing platform already produces a device-use report or whether a new analytics module gets bolted on. For ed-tech vendors selling classroom AI and learning analytics products into the Victorian school market, the buyer's evaluation now includes whether the product can show its time-on-screen footprint to a parent or principal who has to defend the 90 minute cap, and the vendor whose product cannot answer the question is the one the next state procurement form moves down the list.
Data in Schools Australia opens at Melbourne Connect today on whether existing dashboards do the work
The tenth Data in Schools Australia conference opens today and runs through Friday at Melbourne Connect at the University of Melbourne, with the program built around the theme "Make Data Talk: Telling the Story of Student Growth and Impact" and a capped 100 participant attendee list. The published speaker line carries ADSEI founder Dr Linda McIver on data literacy, former ACER chief executive Professor Geoff Masters AO on educational measurement and student growth, and Dr Anne-Marie Chase on AI ethics and equitable data-informed school cultures, with sessions on data interoperability, dashboard design, and evidence-based decision-making. The conference framing names the gap the system has been carrying since the Productivity Commission's interim school agreement work: the school has the data, the school does not yet have the analytical practice that turns the data into a curriculum decision. The Data in Schools Australia 2026 conference page.
Why this matters: For independent and Catholic school principals walking out of the executive meeting where Monday's NAPLAN writing data was tabled, the conference's data interoperability stream is the operational answer the next IT roadmap has to commit to, and the question the data team faces in July is whether the writing data is read against the school's own attendance and behaviour data inside the Sentral or Compass tenant or whether the writing file is sitting in a separate spreadsheet on the deputy's desktop. For state Department of Education data leaders running the cross-platform pipelines behind the SIS to LMS to analytics stack, the McIver and Masters framing is the case file the next budget round has to commit to, and the read the term 3 strategic plan has to answer is whether the dashboard the schools see actually surfaces the patterns the data warehouse is supposed to be holding. For ed-tech vendors selling learning analytics products into the Australian school market (Microsoft Education Insights, Google Education Plus, Tes360, the AERO-aligned platforms), the buyer's evaluation criterion this week is whether the product reads the existing data the school already pays for, or whether the platform asks the school to add another data ingestion step the conference will spend two days naming as the bottleneck.
AARE researcher says GenAI is in Australian childcare without the guidance the school side has
Southern Cross University's Nicole Thompson published a piece in EduResearch Matters on Thursday 4 June arguing that generative AI tools are already being used in Australian early childhood education settings without the sector-wide guidance the K to 12 system has spent the last two years building. Thompson's argument: generic AI tools have arrived faster than the early childhood sector's policy frame can absorb, leaving individual educators making privacy and ethical decisions without the institutional architecture the schools' regulatory frame provides, and the tools' lack of grounding in the Early Years Learning Framework undermines the relationality the EYLF is built around. Her capstone line: "Either the sector shapes GenAI use, or generic tools will quietly shape practice for us." Thompson's specific concerns include young children's inability to consent to information entering AI systems, the silent failure mode where "a learning story sounds fluent but loses the child's voice," and the cultural misalignment with Aboriginal early childhood pedagogies. The EduResearch Matters piece on GenAI in childcare centres, 4 June.
Why this matters: For Catholic and independent early learning centre directors briefing the operations committee against the OAIC Children's Online Privacy Code that registers on 10 December, Thompson's argument is the case file the centre's AI policy paper has to commit to, and the question the next parent meeting answers is whether the families know the AI tools the staff use to draft learning stories. For state and territory regulators running the National Quality Framework against early childhood services (ACECQA and the state authorities), the operational read is that the sector's procurement and policy frame for generative AI sits at least eighteen months behind the school side, and the question the next NQF revision has to answer is whether the standards on relationships, identity, and culture (Quality Areas 1, 5, and 6) actually price the AI tool the service is already using. For ed-tech vendors selling AI products into the Australian early childhood market (the curriculum planning tools, the learning story platforms, the parent communication apps with AI features), Thompson's framing is that the buyer's evaluation column the K to 12 sector added last year is now in the early childhood centre's procurement form too, and the vendor whose product cannot answer the EYLF alignment question is the one the next service decision moves on from.
The Thursday thread is that the device sits inside the classroom under a clock the parents and the state now both read, the data the school already owns is a story the conference asks the system to actually tell, and the early years sector is the one place the procurement frame still moves faster than the policy paper can catch. More on the operational read at digitalattitudes.com.au.