I’m interested in what technology leaves us able to do when we stop using it. A polished answer on a screen is one thing. Being able to explain it tomorrow is another. This week, Australian classroom research added to the digital-detox debate, alongside a larger review of AI in education and proposed rules for frontier models. The question running through them: who is expected to do the work?

Australian schools: what needs detoxing?

The Department of Education’s A Digital Detox paper, dated 4 November, provides the backdrop. It supports purposeful technology use and allows jurisdictions or sectors to introduce, trial or consider screen limits. It is not one uniform national screen ban. [5]

On 9 November, Southern Cross University highlighted findings from the NSW Screens in Education project: purposeful classroom use alongside substantial non-educational use in secondary schools. The government paper cites the same survey project to highlight distraction. These are different aspects of the evidence, not necessarily contradictory results. Surveys and observations cannot establish what a blanket ban would achieve. [7]

The question is what students are doing with the technology, and what they are learning from it.

Most impactful research: finishing isn’t the same as learning

Stanford’s SCALE team published an updated education review on 10 November, expanding its causal evidence base from 20 papers to 52. The distinction that matters: using AI to understand material looks more promising than having it complete the work, although learning gains are inconsistent. [1]

This is a research synthesis, including preprints, with much of the evidence drawn from university settings. It does not establish lasting gains across schools. A student’s satisfaction with the help and what they have actually learnt remain different questions.

For teachers and parents, the finished page is a poor place to stop looking. Ask the student to explain a decision in it. That conversation tells you something the formatting cannot.

Government and policy: responsibility reaches the provider

On 10 November, the Australian Government announced plans for mandatory standards covering frontier AI companies, enforced by a proposed specialist regulator. Independent model evaluation and rapid incident reporting are among the intended requirements. These are announced plans, not a regulator already enforcing a new law. [2]

The same day, the Government released 18 recommendations from its review of an AI agent’s infiltration of a health department website, including the Medicare Statistics Reporting Service Portal. It accepted the recommendations. The incident itself was earlier; the review’s release is this week’s development. [3]

“Use AI responsibly” is a rather incomplete instruction if it applies only to the person at the keyboard.

Something promising: a safer starting point

Ofcom said on 6 November that Snapchat had begun turning off its Find Friends feature for child users by default in the UK. The regulator expects rollout to finish this month, with age checks for under-18 protections expected by year’s end. Neither milestone should be described as complete. [4]

That is a concrete design change worth following. It puts some responsibility into the product, rather than leaving every family to discover which setting needs changing. Whether it reduces harmful contact still needs checking.

The thread through these developments is responsibility. Young people need opportunities to build judgement, and adults need evidence before promising transformation. Providers have work to do too. I’d like us to get better at noticing what a tool asks us to practise, and what it quietly lets us stop practising.

Something worth your attention

Coded Bias, Shalini Kantayya’s 2020 documentary, follows researcher Joy Buolamwini’s investigation of bias in facial recognition. It offers an earlier lens on this week’s accountability questions: who examines a system, and who lives with its errors? It is further viewing, not evidence about current classroom AI. [6]

Sources

[1] Stanford SCALE, 10 November 2026. https://scale.stanford.edu/ai/evidence-base-review/evidence-base-ai-k12-november-review

[2] Australian Government, 10 November 2026. https://www.pm.gov.au/media/new-australian-ai-safety-and-security-protections

[3] Australian Government cyber review release, 10 November 2026. https://www.pm.gov.au/media/strengthening-australia-against-ai-driven-cyber-incidents

[4] Ofcom, 6 November 2026. https://www.ofcom.org.uk/online-safety/protecting-children/ofcom-provides-update-following-its-call-to-action-on-tech-industry-to-improve-protections-for-children

[5] Australian Department of Education, A Digital Detox, dated 4 November 2026 (earlier policy context). https://www.education.gov.au/download/20853/digital-detox/46397/document/pdf

[6] Coded Bias, 2020; official film website and AFI programme. https://www.codedbias.com/

[7] Southern Cross University, 9 November 2026, Do NSW classrooms need a ‘digital detox’? New research challenges common assumptions. https://www.scu.edu.au/news/2026/do-classrooms-need-a-digital-detox/

Film date/director and subject verification: https://www.codedbias.com/filmmaker-1 | https://docs.afi.com/2020/feature-films-2020/coded-bias/