Experts push back on NSW screen time caps

Digital Attitudes, 15 September 2026

Tuesday of Term 3 Week 10 opens, and the sector response to Sunday's NSW screen cap has arrived. Two Australian expert responses in 48 hours, an international counter-example in the same news cycle, and Instructure has quietly named a Chief Privacy Officer to guide its AI product decisions.

120 experts and a CQU neuroscientist ask NSW to measure screen use, not screen time

The University of Wollongong's Dr Karley Beckman and Associate Professor Tiffani Apps, both from the ARC Centre of Excellence for the Digital Child, led a Position Statement now signed by more than 120 Australian experts and organisations (developed with the Alannah and Madeline Foundation) that asks state and federal education ministers to shift from screen time to screen use in schools, and warns that blanket restrictions risk undermining learning, digital literacy and educational equity (EducationHQ on the expert pushback to the NSW screen time cap, University of Wollongong media release with the Beckman and Apps position statement). Beckman, on the record: "what actually matters for learning isn't how many minutes a student spends looking at a screen but what they're doing with it, why, and with whom". The statement notes screen-time guidelines originated as 1990s television advice for toddlers, that neither the WHO nor Australia's own movement guidelines restrict learning-related screen use for children over five, and that Australia recorded its lowest digital literacy levels in 15 years in the same window the sector is talking about caps. CQUniversity educational neuroscientist Professor Ken Purnell gave EducationHQ a separate but adjacent read: "This isn't a battle between pencils and laptops. It is about putting learning first" (EducationHQ on Purnell backing screens reform while warning against missing the bigger picture).

Why this matters: for a NSW deputy principal, director of digital learning, or head of teaching and learning writing a 2027 program on top of Prue Car's Sunday announcement (yesterday's brief) and the NSWEduChat rollout, the response arrived faster than the policy. Two operational reads. First, the political read. On Sunday NSW had a minister, a chief paediatrician, a 28,000 signature parent petition, and the federal education minister publicly backing a digital detox. By Tuesday morning it has 120 experts including two named ARC Centre of Excellence researchers, the Alannah and Madeline Foundation, and CQU's educational neuroscience chair asking the same ministers to walk the numeric cap back. If the NSW 2028 mandatory rule is going to survive the consultation window it now has to answer, on paper, why the Beckman and Apps evidence base does not apply. The Victorian primary device-free rules landing in Term 1 2027 (background: 22 June brief) sit in the same test. Second, the operational read. If a school leadership team is writing the Term 4 procurement paper on Compass, Sentral, Schoolbox, SEQTA, Mathletics, Reading Eggs, Studyladder, IXL, Essential Assessment, or any state-built AI tool, "screen use quality" now has a live measurement question attached. What is the product designed to teach, at what pace, with what teacher agency, and how does the platform report on that against the minute counter the state has proposed. A vendor whose only answer is time on task is now on the wrong side of both the minister's cap and the expert consensus that argues against it. Both readings agree on one thing: the sector data on what students actually do inside a school AI or ed-tech tool is the paper the December 2026 board meeting is going to sit on.

Norway drafts a school GenAI law for six to 19 year olds, and Australia has a comparison to write against

Norway announced in August 2026 that it will legislate age-tiered restrictions on generative AI use in schools, discouraging AI as a general rule for six to 13 year olds, allowing cautious use under teacher supervision for 14 to 16 year olds, and directing 17 to 19 year olds to learn appropriate AI use for further education and work, following the same government's 2024 smartphone ban that the ministry credits with reducing bullying and improving academic outcomes (The Educator on Norway rethinking generative AI in its classrooms). Norwegian Prime Minister Jonas Gahr Støre, on the record: "The most important thing in school is that our children learn to read, write and do mathematics". Legislation is being drafted, funding for classroom books has been increased, and separate restrictions on social media access for children under 16 are being prepared alongside.

Why this matters: for a school council chair, state education department deputy secretary or peak body chief executive reading Sunday's NSW announcement in one hand and today's Beckman and Apps statement in the other, Norway has just handed the sector a live comparison the Term 4 paper has to name. Two operational reads. First, the policy shape. Norway has picked age tiers explicitly for generative AI, not screens generally, and it has written the tiers into legislation rather than into departmental guidance. That is a different design from the NSW minute counter and a different design from the Australian Framework for Generative AI in Schools, which has neither age tiers nor legislative force. If the September or October state council paper wants to argue for a national approach, Norway is the current worked example on the other side of the world from Boston (12 September brief) and New York (7 September brief). Second, the vendor read. State-built AI tools (NSWEduChat for Years 5 to 12, EdChat for South Australian public high schools rolling in from 13 October, Corella in Queensland Catholic, ClassmAIte in Western Australia) have all been designed on the assumption that generative AI belongs in the classroom, with age controls handled at the product layer. A Norwegian-style law would move the age control up to the statute layer and would require every one of those products to prove, on paper, which year levels are permitted to use which features. The Digital Duty of Care exposure draft (9 September brief) closes for submissions at 12pm Tuesday 22 September. If Australia writes an age-tiered rule for classroom AI it will be written in the same window, and the current draft does not yet.

Instructure hires a Chief Privacy Officer to steer AI, four months after the Canvas breach

Instructure announced on Monday 14 September 2026 that it has appointed Stephan Geering as its first Chief Privacy Officer, a role Instructure says will oversee global data privacy, evaluate AI features for safety risks, set data handling requirements for learner information, and guide compliance with emerging international AI regulation, reporting into Chief Executive Steve Daly (Instructure's press release on the Geering appointment as Chief Privacy Officer to guide responsible AI strategy). Geering brings 20 years across technology, financial services and government, and most recently led privacy and AI governance at an ed-tech firm. Daly, on the record: "AI is changing education faster than most institutions can plan for, and getting data privacy and governance right is critical". The appointment lands four months after the May 2026 Canvas breach that Instructure paid ShinyHunters to resolve and that ended up compromising Australian universities and Queensland K to 12 among the 275 million records claimed by the attackers, and the same week UNSW walked away from Moodle and toward D2L Brightspace (Sunday's brief).

Why this matters: for a chief information officer, chief privacy officer, or director of digital learning at any Australian Canvas customer weighing a 2027 renewal against the D2L and Moodle alternatives, this is the vendor answering the exact question the sector spent Term 2 asking. Two operational reads. First, the vendor read. Instructure appointing a Chief Privacy Officer with a specific AI mandate is what the sector board paper says the incumbent should do after a 275 million record breach. The signal to the Melbourne, USyd on the Canvas side, RMIT, Griffith, Adelaide, Newcastle and Swinburne renewal conversations is that Instructure is now hiring against the same competency map D2L just used to win UNSW. The 2027 tender panel that does not ask each shortlisted LMS vendor for a named Chief Privacy Officer with a public AI remit, a security incident review since May, and a written statement on how AI features are cleared for release, is a panel a step behind the market. Second, the framework read. Geering's brief specifically names "evaluate AI features for safety risks" before customer release. That maps to the TEQSA regulatory-led posture for 2026, to the Digital Duty of Care exposure draft (9 September brief) that pulls AI features under the same duty as social media, and to the OAIC Children's Online Privacy Code that must be finalised by 10 December 2026. The Australian LMS or SIS vendor (Compass, Sentral, Schoolbox, SEQTA, Corella) whose 2027 procurement paper does not name an equivalent role with an equivalent brief is now the vendor whose September school council paper has a question mark against it.

Three moves, one Tuesday. The NSW minute counter has met a 120 signature Australian expert consensus that says measure use not time. Norway has drafted a generative AI classroom law with age tiers the Australian framework does not carry. And Instructure has named a Chief Privacy Officer four months after paying the Canvas breach ransom, the week UNSW walked away from Moodle. For a longer read on how to think about the sector's device, vendor, and regulatory floor before Term 4 procurement papers close, see digitalattitudes.com.au.