AI in Schools
ACER takes PISA 2029 AI literacy as ATEC commences 1 July
Digital Attitudes, 10 June 2026
The Wednesday read is the operational paper that lands on Australian education desks this week from four different corners of the same conversation: one new higher ed Commission, one student side AI survey, one classroom feedback paper and one international assessment framework, all setting the term 3 and 2027 procurement frame for what AI literacy in Australian schools and universities actually means.
Glover takes the ATEC desk on 1 July with the Standards Framework
Education Minister Jason Clare and Skills and Training Minister Andrew Giles confirmed in late May that Professor Barney Glover (currently Jobs and Skills Australia Commissioner, formerly Vice-Chancellor of Western Sydney University and Charles Darwin University, and former Chair of Universities Australia) takes the inaugural Chief Commissioner seat at the Australian Tertiary Education Commission for a five year term commencing 1 July 2026. The Hon Fiona Nash, David Coltman and Professor Stephen Duckett take statutory Commissioner seats for three year terms from the same date. ATEC was formally established on 29 April 2026 after the enabling legislation passed Parliament on 31 March, and the body now takes responsibility for the Higher Education Standards Framework from the existing Higher Education Standards Panel, allocates funding under the new Managed Growth Funding system, and negotiates mission based compacts with the 39 universities. Clare's line on the structural goal: "A system that's more seamless and more connected, where it's easier to move between TAFE and uni." Universities Australia Chief Executive Luke Sheehy's caution: against adding "another layer" of oversight to an already heavily regulated sector. The Pie News on the Glover appointment, 1 June, and the National Tribune carriage of the ATEC formal establishment release, 29 April.
Why this matters: For vice-chancellors and deputy vice-chancellors (academic) reading the institution's AI assessment policy against the TEQSA toolkit update of 8 May and the Universities Australia 15 May TEQSA response together, the ATEC commencement on 1 July is the date the Higher Education Standards Framework moves to a new desk, and the Framework is the document the institution's own AI assessment redesign sits under. For TAFE NSW, TAFE Queensland, TAFE SA and TAFE WA directors reading the Glover background (the Jobs and Skills Australia Commissioner seat that ATEC effectively inherits the perspective from), the operational read is that the VET to higher ed AI tooling conversation now runs through one Commission rather than two, which is the point the SupaHuman and EduGrowth VET working group covered on Friday was already arguing for. For ed-tech vendors selling AI tools, assessment platforms and academic integrity products into Australian universities, the procurement frame to write the next contract against is that the regulator stack now reads TEQSA (compliance) plus ATEC (Standards Framework and funding compacts), which is a different audit pattern than the same vendor pitched against in the 2025 renewal cycle.
Anyway, KPMG and Microsoft survey lands 10 percent youth AI job confidence
Career platform Anyway, KPMG Australia and Microsoft published Young Australians and the AI Workforce Transition on 26 May, drawing on responses from 1,029 Australians aged 15 to 24. The headline numbers: 45 per cent of respondents now report a strong understanding of AI (up from 27 per cent in 2023), 41 per cent use AI for learning and assignments, 15 per cent use AI for work or career development, only 10 per cent believe AI will help them land a job, and 63 per cent expect AI to eliminate roles. 27 per cent of respondents say schools or universities actively discourage AI use. The geographic and socioeconomic split: rural respondents (47 per cent non-use) and low socioeconomic respondents (52 per cent non-use) trail higher socioeconomic respondents (35 per cent non-use) on AI engagement. Anyway co-founder Will Stubley's line: "If we can bridge that and give young people the tools and guidance to turn those skills into something practical, that's where the real opportunity is." The Educator K/12 read of the survey, 27 May, and the HR Leader report on the survey findings, 25 May.
Why this matters: For independent and Catholic school heads briefing the term 3 careers and AI literacy program this week, the 10 per cent confidence figure is the line every careers conversation will be read against, and the 27 per cent "school discourages AI" data point is the line every school's published AI policy will be measured against by parents at the next information evening. For state Department of Education CIOs running the careers stream alongside the NSWEduChat, SA EdChat, Corella and ClassmAIte deployments, the rural and low socioeconomic non-use figures are the equity read the procurement team will hand back when the next deployment paper lands, because the answer to "why is engagement uneven by postcode" is not the licence cost, it is the school's careers and pastoral plan. For TAFE directors and university deputy vice-chancellors (academic) writing the graduate employability response to the Senate failed promise inquiry that closed Friday, the survey is the cleanest student side dataset the response can cite, because the inquiry's framing is the student does not trust the degree to certify the work, and the survey says the student does not trust AI to deliver the job either.
ACER and Melbourne argue AI feedback that fixes work misses the point
Dr Fabienne van der Kleij (Senior Research Fellow in Assessment Transformation at the Australian Council for Educational Research) and Professor Therese Hopfenbeck (Director of the University of Melbourne's Assessment and Evaluation Research Centre) published in Teacher Magazine on 3 June a research backed argument that AI feedback tools in classrooms have to be designed to improve student learning and self-regulation, not to correct student work. The piece names three misconceptions the AI feedback procurement question lands on (feedback effectiveness depends on message features alone, more feedback always improves outcomes, students automatically understand and apply feedback) and three student capabilities the tool has to develop (evaluative judgement, feedback seeking, agency and ownership). The line that names the operational test: "If feedback focuses on fixing students' work for them, then we're missing the point." The van der Kleij and Hopfenbeck piece in Teacher Magazine, 3 June.
Why this matters: For independent and Catholic school heads procuring the next AI marking or AI feedback product into the secondary marking workflow, the van der Kleij and Hopfenbeck question is the one the procurement team has to hand the vendor: does this tool make the student a better self-regulator, or does it make the work submit ready. For state Department of Education CIOs running the NSWEduChat student mode, the SA EdChat secondary deployment and the Corella Year 7 to 10 access, the piece is the academic read that backs the student mode design choice (guided prompts rather than direct answers) and the line every parent consent form will read alongside. For ed-tech vendors selling AI marking, AI feedback and AI tutoring products into Australian classrooms (the Eduaide and MagicSchool suites, the Microsoft Teach for educators tool, the Education Perfect AI marker, the Stile AI Library), the operational read is that the buyer side now has an academic baseline for what an acceptable AI feedback product looks like, and the standard line the buyer will run is the van der Kleij quote against the product demo.
ACER takes PISA 2029 and runs the first international AI literacy test
The OECD has appointed the Australian Council for Educational Research to lead PISA 2029 across 90 plus countries, with Media and Artificial Intelligence Literacy (MAIL) as the innovative domain for the tenth cycle, the first international comparable assessment of AI literacy at scale. The MAIL Assessment Framework first draft, published by the OECD in January, runs the assessment across three dimensions (authors and audiences, messages and meanings, representations and realities) and maps the framework against eight subject areas (Reading, Mathematics, Science, History, Geography, Foreign Languages, the Arts, Computer Science). Field trials run in 2028 with the main survey in 2029. ACER CEO Lisa Rodgers's line: the 2029 cycle arrives "at a critical juncture" for assessing student capabilities with AI. PISA 2025 results land 8 September 2026. Teacher Magazine on the MAIL framework and curriculum links, 29 April, and the ACER release on the OECD appointment, 1 April.
Why this matters: For independent and Catholic school heads briefing the 2027 to 2029 curriculum and assessment review, the MAIL framework is the document the Year 6 cohort this year will sit against four years from now, because the 15 year olds in the 2029 cycle are this year's Year 6 students, and the school's curriculum redesign has a four year runway to make the literacy line a teachable competence. For state Department of Education CIOs reading the AITSL Teacher Standards review (open through to February 2027) alongside the MAIL framework, the operational read is that the AI competence question the AITSL submission has to answer is also the question the international comparable assessment will measure four years later, which means the Standards line and the assessment line need to be drafted as one document. For ed-tech vendors selling AI literacy modules and digital citizenship curriculum into Australian schools (the Common Sense Education AI Foundations module, the various Microsoft Elevate AI Literacy for Educators credentials, the Atlassian Foundation AI literacy programs), the MAIL framework's three dimensions are now the public template the procurement team will run the product specification sheet against.
The thread this morning is that the operational paper for AI literacy in Australian education is now landing from four desks at once. Glover's ATEC takes the Higher Education Standards Framework on 1 July, Anyway lands the 10 per cent youth job confidence figure, van der Kleij and Hopfenbeck put the academic read on AI feedback design, and ACER takes PISA 2029 with the first international AI literacy assessment. More on the operational read at digitalattitudes.com.au.