I grew up on devices. That is not a confession — it is just a fact, and I am mentioning it upfront because this post is going to raise concerns about children and technology, and it would be dishonest to pretend I'm writing from some screen-free pastoral idyll. I had a Nintendo DS. I had YouTube. I had Minecraft at a formative age and I am not sorry about it. What I also had, which I increasingly suspect is the thing that made the difference, was a substantial chunk of life that happened offline. Boredom. Unstructured time. Books that I finished rather than excerpts designed to hold attention for fifteen seconds before the algorithm moved on. A childhood that technology was part of, rather than a childhood that technology was.
The two generations coming up behind me do not have that. And I think we are only beginning to understand what that actually means.
The iPad Kid Is Not a Joke Anymore
The term "iPad kid" started as internet shorthand — Gen Z side-eyeing the toddlers of millennial parents, glued to a screen in a restaurant while Cocomelon played at full volume. It was funny in the way a lot of generational observations are funny: enough truth to sting, wrapped in enough irony that you could pretend you were just having a laugh.
Studies have shown that excessive screen time at a young age can be harmful and lead to struggles with attention span, emotional control, and social interaction — which is often what people are describing when they talk about iPad kids. That is not a culture war claim. That is the developmental psychology research. A 2025 report by Common Sense Media found that 40% of children have an iPad by the time they are two years old. Two. Before many of them can form complete sentences, they have a device optimised by an army of engineers to hold their attention as long as possible.
Children aged 8–12 in the US now average around four hours and forty-four minutes of daily screen media. Teenagers average seven hours and twenty-two minutes. That is not occasional use. That is a primary medium. And the COVID-19 pandemic accelerated all of this considerably — lockdowns and the shift to online education caused a significant increase in electronic usage in young children at exactly the age when offline interaction, physical play, and unstructured time are most developmentally important.
Generation Alpha — born between 2010 and 2025, the first generation to be born entirely in the 21st century — is often characterised by their unprecedented reliance on technology. That framing is usually presented neutrally, as if "unprecedented reliance" is simply a descriptor of the times rather than something worth examining. I am less relaxed about it than that.
I also have a fairly direct view of what it looks like in practice. I have twin nephews on the earlier end of Generation Alpha who were raised in a way that rhymes with how I grew up — a reasonable amount of offline time, the kind of upbringing where boredom was permitted to exist and resolve itself without a screen intervening. When I see them, they are happy enough to be somewhere without a device in hand. They are, from what I can observe, sociable. Then there are nieces and nephews further into the generation — the late Alpha end — who are, and I say this without any intended condemnation of them or their parents, textbook iPad kids. I do not see any of them particularly often, so I am working from a limited sample, but the contrast is difficult to miss. The earlier set will make conversation. The later set have a noticeably different relationship with unstructured time. Whether that is down to the devices specifically, or to wider shifts in how that cohort is being raised, I cannot say with certainty. Probably both. But the difference is there, and it maps neatly onto the data.
What the Literacy Numbers Actually Look Like
Here is where it stops being abstract.
According to the Annie E. Casey Foundation, only 30% of Generation Alpha fourth graders read proficiently as of 2024 — down from 32% in 2022. That is a decline over two years in a metric that has been declining for longer than two years. In the most recent Nation's Report Card data, 40% of fourth graders, 33% of eighth graders, and 32% of high school seniors were reading below the Basic reading level.
The explanations for this are not simple, and I am not going to pretend they are. Curriculum decisions — specifically the widespread adoption of Whole Language reading instruction over phonics — contributed to this independently of screen time. But technology is not an innocent bystander. The devices children use for reading are often also filled with apps, games, and social media platforms designed to capture their attention — and that fragmented focus makes sustained engagement with long-form text genuinely difficult to develop.
Novels have been removed from most schools' curricula and replaced with short excerpts, partly to accommodate reduced attention spans. Which is, if you think about it, the educational system accommodating the problem rather than addressing it. You cannot build reading comprehension on a diet of content designed to be consumed in fifteen-second segments.
The spelling angle gets discussed less, possibly because "kids can't spell" sounds like every generation complaining about the next one. But the mechanism here is specific and worth naming. Autocorrect and predictive text mean that the cognitive loop of encoding a word — thinking about its structure, making an attempt, self-correcting — simply does not happen. The device does it. Over years of use beginning before literacy is properly established, the skill does not develop because it never needed to. I have seen this in practice, and not just anecdotally: teachers have been going viral on TikTok — on the very platform accelerating the problem — reporting that some children cannot properly spell their own names. That is not moral panic. That is primary school teachers, in significant numbers, saying the same thing independently.
Generative AI is autocorrect taken to its logical conclusion. If autocorrect meant you never had to think about how a word was spelled, AI means you never have to produce the sentence at all. By May 2025, 84% of US high schoolers were using generative AI for schoolwork — and that figure is almost certainly higher now and almost certainly undercounts students who do not identify what they are doing as "using AI" because it has become invisible infrastructure. The concern is not academic dishonesty in the narrow sense. The concern is that writing is thinking. The process of forming an argument in prose, finding the words, deciding what order things go in — that is not a neutral transcription of pre-formed thought. It is how the thought becomes coherent in the first place. Outsource it entirely and you do not just produce worse essays. You practice reasoning less. For a generation already contending with shortened attention spans and a platform environment hostile to nuance, AI completing their written work is not a shortcut. It is the removal of one of the few remaining cognitive workouts they were getting.
TikTok's contribution to this is worth its own paragraph. Research has found that the platform's fragmented, fast-paced content diminishes attention spans and fosters addictive behaviours among its youngest audience — and the subtle nature of influencer-driven content complicates children's ability to distinguish advertising from genuine communication. The written language on TikTok has its own evolving register — abbreviations, neologisms, phonetic spelling used for comedic or in-group effect — which is fine as a cultural register and actively harmful as a primary model for literacy development in a child who has not yet established baseline competency.
I have been on TikTok for a few weeks. I want to be upfront about that, because I am aware it makes me a fairly new arrival making observations about a culture I have not had long to absorb. But sometimes the fresh pair of eyes is the useful one. The comment sections are, and I am trying to be precise rather than dismissive here, frequently extraordinary. Not in the good way. The combination of phonetic approximations, context-free aggression, and apparent incomprehension of whatever is being commented on produces a specific kind of written register that I think has to be seen to be fully appreciated. What strikes me more than the spelling, though, is the near-total absence of critical thinking. Not nuanced critical thinking — just any. The response to almost anything is immediate, binary, and completely unexamined. That is not a coincidence. That is what you get when a platform is engineered to produce dopamine hits in rapid succession from early childhood: you do not develop the habit of pausing between stimulus and response, because the entire design works against it. What you get instead is dopamine burnout dressed up as engagement — a generation that has been conditioned to react rather than think, and that will carry that reflex into every context it encounters.
Then There Is Generation Beta
Generation Alpha at least grew up watching the technology arrive. They were there when the iPhone became ubiquitous, when TikTok displaced everything else, when the smart speaker appeared on the kitchen counter. They experienced the before, however briefly.
Generation Beta — born from 2025 onward — will be the first cohort born entirely into a world dominated by artificial intelligence and virtual reality. Unlike Millennials or Gen Z, who learned to adapt to AI as it emerged, Generation Beta will not know a world without it. The implications of that are genuinely difficult to think through clearly, because we are talking about children who will have no experiential baseline for what human cognition unassisted by AI actually looks like. Digital identity management will begin early, with data privacy, cybersecurity, and online safety becoming central parenting and societal concerns from the moment these children first engage with technology — which, on current trends, will be very early indeed.
Think about what that actually means developmentally. Generation Alpha grew up outsourcing spelling to autocorrect and attention to algorithms. Generation Beta will grow up outsourcing reasoning itself to AI. Not as a tool they reach for when stuck, but as the default first response to any cognitive demand. Why work through a problem when the assistant in your ear will work through it for you? Why form an opinion when you can ask something to form one? The risk is not that they will be less intelligent. The risk is that they will never build the habit of thinking independently, because the scaffolding was always there and they never had to construct it themselves. We already have some evidence of this pattern emerging in older students; in children who have never known the alternative, the effects are likely to be considerably more pronounced.
There is also the question of what it means for their relationship with truth. AI systems generate plausible-sounding output. They do not reliably generate accurate output, and they are getting better at the former faster than they are getting better at the latter. A generation raised to treat AI responses as authoritative — because the device has always been authoritative, because the iPad always had the answer — is not going to develop strong instincts for verification. They will read a confident, well-formatted AI response and have no particular reason to question it, because questioning the device is not a reflex they were ever given the opportunity to form.
The Cybersecurity Problem Nobody Is Teaching
I have just spent two years studying OCR Cambridge Technicals in IT Level 3, with a focus on application development. Cybersecurity was one of my three exam units — not a passing module, but a subject I sat an exam in and spent a meaningful amount of time working through in depth. This is the part of the post where I stop being a concerned observer and start being genuinely alarmed, because the cybersecurity picture is, in my view, the most serious dimension of all of this and the one that receives the least attention in public conversation about children and technology. And AI, for everything it promises, is making that picture considerably worse.
What the literacy decline costs you is legible and, at least partially, recoverable. What the security gap costs you can be immediate, severe, and in some cases permanent — and the generations walking into it have been given almost no tools to navigate it.
The core problem is this. Fluency with the interface breeds the assumption that you understand the system beneath it. You do not. Knowing how to navigate TikTok's UI tells you nothing about how the platform handles your data, nothing about how to evaluate a suspicious link, and nothing about what happens when your account credentials appear in a breach database. These are different skills entirely, and one does not imply the other. This is the digital native paradox, and it is actively dangerous.
Research shows that while Gen Z quickly adapts to new tools, they are more likely to fall for phishing scams than older generations — a mix of confidence and complacency that makes them both valuable and vulnerable. That finding should be alarming on its own. It becomes more alarming when you consider what it implies for the generations behind them, who will have even less friction in their relationship with technology and therefore even less occasion to develop a healthy scepticism of it.
The specific behaviours are worth naming rather than gesturing at vaguely. A 2025 report found that common risky behaviours among digital natives include using weak passwords such as birthdays and pet names, connecting to public Wi-Fi without a VPN, and sharing sensitive work or personal information with AI tools. The same report stated that 56% of Generation Z digital natives have had no access to any cybersecurity training and are ill-equipped to identify and prevent attacks. 56%. That is not a marginal group. That is the majority of the generation that is now entering the workforce and will shortly be raising Generation Beta.
And that is just the known surface. The attack surface these generations are actually operating on is considerably wider.
Credential stuffing — automated attacks that take breached username and password combinations from one service and try them against others — is only a threat if you reuse passwords. Password reuse is near-universal in people who have never been told why it is dangerous. A child who has been handed an iPad at two years old and given no accompanying security education is going to grow up creating accounts with the same email and the same password, for the same reason they will use their birthday as a PIN: because nobody told them not to, and the device made it easy.
Social engineering and phishing have also become considerably more sophisticated in the time it has taken Generation Alpha to grow up. The obvious phishing email — broken English, improbable urgency, a link that looks wrong if you squint at it — is no longer the primary vector. AI-generated phishing is personalised, grammatically correct, and contextually plausible. It references your name, your employer, your recent activity. For a generation that has grown up sharing everything online and has an extensive, scraped-together data trail from early childhood, the targeting becomes trivially easy. You do not need to trick someone who has told the internet everything about themselves since before they could consent to doing so.
Deepfakes extend this further. AI-generated audio and video of real people — sufficiently convincing to deceive someone who is not specifically looking for signs of manipulation — are no longer expensive or technically complex to produce. A child who has been taught to trust the face and voice on the screen, which is to say every child raised on YouTube and TikTok, has no particular defence against a video of someone they recognise asking them to do something. The obvious examples are financial fraud and social engineering, but the same technology works for coercion, for disinformation, for manufacturing apparent evidence of things that never happened. The generation most exposed to synthetic media is the generation least equipped to identify it, and that combination is genuinely frightening to think about at scale.
Then there is the data broker problem, which almost nobody talks about with children. Data brokers are companies that aggregate personal information — names, addresses, phone numbers, browsing habits, purchasing history, location data — and sell it. They are largely unregulated in the UK and most other jurisdictions. A child who has spent their entire childhood on devices they do not own, using platforms with terms of service they did not read and could not have understood, has been building a commercial profile that will follow them into adulthood. Generation Beta will begin accruing this profile before they can speak. The data will outlive their childhood by decades. Most of them will never know it exists.
SIM swapping — where an attacker convinces a mobile network to transfer your number to a SIM they control, bypassing SMS-based two-factor authentication — is increasingly common and disproportionately affects people whose personal details are easily available online. Again: a generation whose details have been publicly available since infancy, who has never been taught what two-factor authentication is or why the SMS implementation of it is specifically weak, is not well-positioned to defend against this. It is not a question of intelligence. It is a question of not having been given the relevant information.
Gen Z is already described as having a strong sense of apathy about these problems, believing there is nothing that can be done about them at this point. That apathy is not irrational given the information environment they grew up in — data breaches are so common that they are treated as weather. But apathy about a threat does not make you less exposed to it. It makes you more exposed to it, because you have stopped thinking about the ways you might reduce the risk.
Generation Alpha and Beta are going to inherit all of this at greater scale, with more data exposed, more AI systems involved, more services integrated, and a digital identity that has existed since before the person it belongs to was old enough to understand what that means. This constant connectivity exacerbates risks including cyberbullying, misinformation, and exposure to inappropriate content. Those are the visible ones. The less visible ones — the credential stuffing, the SIM swapping, the AI-targeted phishing, the data broker profiles, the long tail of a scraped-together childhood — are harder to see and more difficult to address, and there is currently no meaningful, mandatory programme in UK schools that addresses any of them seriously.
Children are taught to use software. They are not taught to think critically about it, to understand even the basic model of how their data moves, or to recognise the difference between a service and a product they are. The distinction matters enormously. When a service is free, the business model is usually you. A child who has grown up using Google, TikTok, Instagram, and YouTube without ever being told that is not digitally literate. They are just a user — which is precisely what those platforms need them to remain.
I find that deeply concerning. Not in the abstract, policy-document sense. In the sense that the people I am describing include children I know, and a generation that will be running infrastructure, handling sensitive data, and making security decisions in organisations within the next decade, with a foundational gap that nobody is currently treating as the emergency it is.
The Balance I Had and They Might Not
I want to be careful not to sentimentalise this into "things were better when we played outside." They were not simply better. I had access to online spaces that were also poorly moderated, also capable of eating hours that could have been used differently. I used to gawk at gore online. In retrospect, that did some things to my psyche that I am still not entirely sure how to account for. The internet was not a safe, well-curated environment when I was growing up in it, and pretending otherwise to make a generational argument would be dishonest. It did not suddenly become problematic when Generation Alpha arrived.
What I did have, though, was the before. I know what it feels like to be bored without a screen available. I know what it is to read back to back — space books, biology books, entire series consumed in sequence because the next one was simply there and I wanted to know what happened. Horrible Histories. The World of Norm. I still do it occasionally; I read all twenty-two volumes of BEASTARS — 196 chapters — in six hours, and I am slowly working through Brigid Ehrmantraut's Celtic Magic: A Practitioner's Guide at the other end of the pace spectrum. The point is not that these were highbrow choices. The point is that the sustained attention required to follow a narrative across a hundred and fifty pages, or across a series, is a different cognitive mode from the one TikTok is training — and I had years of practising it before I had a device in my pocket that offered an easier alternative. I know how to spell because autocorrect did not exist in a useful form until I had already spent years writing by hand and on keyboards where I had to produce the letters myself. That scaffolding — the offline foundation on which the online was built — is not something Generation Alpha had access to in the same way, and Generation Beta will have even less of it.
That quote is from a few years ago now. We are starting to get some data on how it is turning out. Declining reading proficiency. Teachers reporting fundamental literacy gaps. A generation that is, as a majority, untrained in cybersecurity and walking into an attack surface that grows more sophisticated by the month. A generation behind that will be born into AI as ambient infrastructure, with no concept of what a world before it felt like and a data trail that predates their first memory.
I do not think the answer is banning iPads or closing down TikTok or pretending that technology is inherently harmful. It is not. AI is not inherently harmful either — I use it, I build with it, I think it is genuinely interesting. But I am no longer convinced that "we need better digital literacy education" is a sufficient response to the scale of what is accumulating. That framing implies a manageable gap to be filled with the right curriculum. What I actually think is that we are watching two generations grow up with a structural deficit — in literacy, in critical thinking, in independent reasoning, in security awareness — that is being treated as somebody else's problem to solve later, and that "later" is running out. AI did not create that deficit, but it is accelerating it in every dimension simultaneously: faster erosion of writing ability, faster erosion of critical thinking, more sophisticated attacks, more convincing disinformation, and a generation being primed from infancy to trust the output of systems they will never understand.
I learned most of what I know about how any of this works by being curious enough to go and look into it myself. That should not be the only route available. And for the generations growing up now, the cost of not knowing is considerably higher than it was when I was figuring it out. The threats are more sophisticated, the data exposure is greater, and the window between "doesn't know about this" and "has already been compromised because of it" is narrower than it has ever been.
That is what keeps me up about it.