This week: children’s views on AI inevitability; parents and children confirming the Online Safety Act isn’t working as intended; the wide cultural divide on AI optimism.
🔍 This week I’ve been thinking about…
The heaviest AI users in Gen Z — the ones on it every day — showed an 18-point drop in excitement about it over the past year. This is not how the typical hype-cycle goes. AI is doing the opposite than expected here.
This comes from a Gallup poll. Across the full Gen Z sample, excitement is down 14 points. Hope is down 9; anger up 9. Eighty percent of Gen Z say AI is likely to make future learning harder. In the same survey, the proportion who believe they will be prepared to use AI after high school rose 12 points in a year, to over 50 percent.
Those two numbers, sitting next to each other, describe something that is not confusion and not optimism. It is adaptation.
The teenagers are absorbing two positions at once: AI is hurting their learning, and AI is what they believe they need to engage with for their future. They have not resolved this contradiction because the adults around them have not resolved it.
📰 Three things worth your attention
1. Kids Can Feel AI Hurting Them, They Have to Use It Anyway — Psychology Today
This is the article that drew my attention to the Gallup poll data. What looks like teenage contradiction — AI is hurting me, AI is my future — is better understood as cognitive dissonance: children caught between an environment that compels AI use and their own clear-eyed observation that it is working against them. I wouldn’t normally put in a piece from Psychology Today, but the reflection on what living with this cognitive dissonance does is helpful to reflect upon.
Children build motivation by thinking that their own judgment shapes the world they live in. In this case, there is a risk that the message about the need for AI — the message being pushed out by the adults with power in their lives — is at odds with their lived existence. When you accept that the thing that is important is not the thing you feel and experience, you can lose the message that you can effect change; that your views matter. Given the ongoing questions as to how much children’s views are being listened to at present, this is quite a depressing thing to think about.
2. The Online Safety Act: Are children safer online? — Internet Matters / The Register
Internet Matters has published the first substantial evaluation of the Online Safety Act from the perspective of families, drawing on a survey of 1,270 children and parents alongside focus groups conducted earlier this year.
Around seven in ten children and parents have noticed new safety features — better reporting tools, content filters, age checks — and most of them welcome the changes. Optimism is cautious but present: 42% of children feel the online world has become safer for them recently.
What the report is more useful for, though, is what it shows is not working and what is not being addressed at all.
On age verification: 46% of children believe age checks are easy to bypass, and 32% say they have done so in the past two months. The methods range from entering a false birthdate to submitting video footage of another person’s face — or, in one case reported by a parent, and picked up prominently by The Register’s article, a child drawing a moustache on themselves with an eyebrow pencil, which the system accepted. More striking still: a quarter of parents have allowed their child to bypass age checks, with 17% actively helping them do so. That cost-benefit analysis is helpful to consider: examples given included games parents had played and felt were appropriate for the maturity of their child, and allowing children to make videos in situations where the audience was known.
On scope: the issues families describe as their most immediate concerns — the amount of time children spend online, and the growing presence of AI-generated content in their feeds — are almost entirely outside what the OSA addresses. Mixed views were reported on the idea of a social media ban, with those against it suggesting it would be ineffective or detrimental. Stronger enforcement of the OSA, stricter age-checks and restricting harmful features were the most popular alternatives to a ban.
3. AI optimism surges in Asia, unlike in the US — Rest of World / Stanford HAI
A Stanford University study finds a striking divergence in how different populations relate to AI. In the US, 38% of respondents describe AI as making them excited. In China, that figure is 84%. In Indonesia, Thailand, and Malaysia, it sits between 77 and 80%. Trust in government to regulate AI responsibly: 31% in the US, 81% in Singapore.
The Rest of World framing is primarily about innovation ecosystems and where AI talent is concentrating. But the finding matters differently for this newsletter’s concerns. Children growing up in countries with high AI optimism and high institutional trust are inhabiting a fundamentally different relationship with the technology than children growing up in the UK or US, where adult anxiety is ambient, political messaging is contradictory, and policy is reactive. The cognitive dissonance described in the Psychology Today piece above may not be a universal experience of adolescence in the AI era. It may be a specifically Western one — shaped by the particular context in which these children are encountering the technology.
CAISE’s policy and media reviews will draw on Chinese policy and media coverage alongside Western sources precisely because of this. The question of what AI does to children’s development cannot be answered without asking: whose childhood, in what institutional context, under what regulatory assumptions, at what moment in their country’s relationship with the technology.
🔁 ICYMI
London schools trialling VR to relieve pupils’ stress — The Guardian
All 15 secondary schools in the London borough of Sutton are using VR headsets in a pilot alongside the local NHS mental health trust. The programme, which is seven minutes long, is used when pupils become overwhelmed in class: by exam anxiety, ADHD, difficult mornings, things that happened at home. Nine out of ten pupils who used the headsets in the first ten schools saw an immediate drop in stress. Schools report reductions in pupils needing to leave lessons.
The honest context for this is that CAMHS services across England are severely overloaded. VR as a low-cost school-based intervention is partly a response to a service that cannot cope with demand. That is not straightforwardly good news, even if the outcomes are promising.
But the design logic is worth noting. The programme was built around what overwhelmed children actually need in an acute moment: something that breaks the loop, quickly, without requiring a waiting list or a referral. It is designed to meet the child’s need, and the child is the one that asks to use it. It seems to be a solution supporting both the child and the system, which is lovely to read about!
🔬 What’s new with CAISE
More logistics, but also the start of some fieldwork! It’s very exciting to get started.
→ What are you seeing in your school, your research, or your own use of AI this week?
Let me know, or share this with someone who is trying to figure it out.