AI in Schools
NAP-ICTL drop, NAPLAN reform call, KPMG AI skills gap
Three reports in 48 hours all point at the same gap: Australian students use plenty of digital tools, but the demonstrated skill, the assessment that picks it up, and the career confidence to use it are all moving the wrong way.
NAP-ICTL drops to a 21-year low and the device spend is the policy question
ACARA published the 2025 NAP-ICTL results this week. Year 6 proficiency fell to 50 per cent (down from 55 per cent in 2022) and Year 10 fell to 37 per cent (down from 46 per cent in 2022, and the lowest result since the assessment began in 2005). The sample covered 5,498 Year 6 students and 4,753 Year 10 students on tasks like creating digital presentations, analysing data, designing algorithms, and responding to online safety scenarios. Regional and remote schools lagged metropolitan ones, and Indigenous students reached the proficient standard at 23 per cent (Year 6) and 19 per cent (Year 10). The Conversation, 26 May, on the Year 10 result.
Why this matters: This is the first NAP-ICTL result since the post-pandemic device boom, and the answer to the question every IT lead has been quietly asking, whether two decades of classroom device spend produces better digital learners, is now in the data and the answer is no. For state Department of Education CIOs and for independent school business managers signing the 2027 device cycle paper, the procurement question shifts. It is no longer how many devices per student, but which device-paired curriculum and which teacher training plan produces a Year 10 student who can actually analyse a dataset, not just open the file. The regional and Indigenous gaps are the more uncomfortable signal: the schools with the strongest 1:1 fleets are not closing the gap, which suggests the platform itself is not the lever. For ed-tech vendors selling adaptive platforms (Mathletics, LiteracyPlanet, Education Perfect, the various PAT-aligned products), the next tender response will need to engage the NAP-ICTL evidence directly. The argument that more screen time produces more skill just failed its largest Australian field test.
AGPPA tells Canberra the NAPLAN writing test should go, replaced with adaptive AI-aware tasks
The Australian Government Primary Principals Association, representing 7,000 principals across 5,300 government primary schools, met in Canberra this week and called for the Year 5 NAPLAN writing test to be abolished. President Pat Murphy cited the widespread disruptions of the March 2026 NAPLAN window (the platform outage that locked roughly one million students out on the morning of 11 March) and questioned the writing component's fairness, consistency, and validity. The AGPPA's proposed replacement is adaptive online testing, sample-based assessment, shorter targeted tasks, and a model that lets students use digital and AI tools while being assessed on creativity and critical thinking. Murphy: "A modern assessment program should assess how students think, communicate, and apply judgment in an AI-rich world." The Educator on the AGPPA position, 27 May.
Why this matters: This is the first sector body to put a concrete replacement model on the table after the March outage, and it lands two weeks out from the Education Ministers Meeting that will set NAPLAN's 2027 direction. For state ministers reading the AGPPA position alongside ACARA's NAP-ICTL slide, the operational tension is visible: the writing test is the only NAPLAN component that produces a sample of student composition (the others are multiple choice), and removing it means losing the signal at the moment digital writing tools are most reshaping student work. The AGPPA's argument is that the signal is already corrupted by the March platform failure and the AI question, and the cleanest fix is a redesign that prices AI in rather than around it. For ed-tech vendors building adaptive assessment products (Education Perfect, Edrolo, and Stile each have an angle), the AGPPA position is the procurement door opening. For Janison and the ESA platform vendors who carried the March outage, the same position is the cover for a tougher conversation with the sector body that depends on their platform working in March 2027.
KPMG and Microsoft find young Australians have the AI use, not the AI confidence
The Anyway, KPMG Australia, and Microsoft Generation AI report released this week surveyed 1,029 Australians aged 15 to 24 and found 41 per cent already use AI for assignments and learning, but only 15 per cent use it for work or career development. Only 33 per cent feel confident explaining AI use in a job interview, 27 per cent say their schools or universities actively discourage AI use, and 63 per cent expect AI to eliminate jobs against 35 per cent who expect equivalent job creation. KPMG's John Munnelly, Microsoft's Tim Allen, and Anyway's Natasha Markham frame AI capability as a foundational skill that should be embedded across subjects rather than confined to the IT class. The Educator on the Generation AI report, 27 May.
Why this matters: For senior secondary heads of department reading this alongside the NAP-ICTL drop and the AGPPA call to bring AI into NAPLAN, the three documents say the same thing from different angles: the students are using the tools, the schools are not yet teaching them what to do with the tools, and the data shows the gap. For careers and pathways staff, the 27 per cent "schools discourage AI" figure is the operational red flag, because an institutional policy designed to protect assessment integrity is producing a graduate cohort that cannot articulate AI use in the room where a hiring decision happens. The Microsoft Elevate for Educators credential, the NSWEduChat and SA EdChat rollouts, and the Brisbane Catholic Copilot for students program are all answers to part of this gap, but the harder operational question is whether the school's actual practice (the lesson, the assessment, the conversation with the careers counsellor) builds the confidence the survey is finding missing. The procurement angle is sharper for the vendor side: a product that does not let teachers run the "show me how you used it" conversation in class is a product that fails the KPMG test.
The thread this morning is the gap between what Australian students are doing with digital tools and what they can actually demonstrate, in three different reports in 48 hours. The schools whose 2027 papers price the gap will be the ones whose digital learning numbers move; the schools that treat NAP-ICTL as ACARA's problem will see the gap widen. More on the operational read at digitalattitudes.com.au.