[Panel sharing at the event Parenting in the AI Age on March 20, 2026]

AI is rapidly becoming one of the most powerful general‑purpose technologies humanity has ever built, reshaping how we consume information, entertain ourselves and relate to one another. It offers phenomenal benefits, but it also stress tests our minds. If we are not careful, AI will not just make information abundant; it will make attention scarce and thinking optional. That is how we drift into artificial ignorance: a state in which powerful tools do so much of the visible thinking that we still look intelligent on the surface, while the underlying muscles of attention, memory and judgment quietly atrophy. This is the real risk facing our children, and those raising them. The question is not only whether AI will grow more intelligent, but whether we will allow ourselves to grow less so.

1. Innovation outruns adaptation

For the first time, the rate of innovation feels consistently faster than the rate of adaptation. We barely absorbed GPT‑3 in 2022 before more capable models landed in 2023 and 2024. Agentic systems now act as digital staff, planning and coordinating quietly in the background.

The curve of technological change has risen above the curve of human or organizational adaptation. Scott Brinker’s Martec’s Law puts it more formally: technology changes exponentially, organisations change logarithmically1 [1].

The gap between those curves is where parents and educators now live. Children inhabit a world of ambient, on‑demand intelligence; adults are still updating policies and habits designed for a slower era. Nowhere is this gap more visible than in how we consume information and spend attention.

2. Information overload with synthetic content

Analysts now warn that we are racing toward a world where much, if not most, online content is synthetic. A Europol‑linked briefing once estimated that “as much as 90% of online content may be synthetically generated by 2026”2 [2]. The precise number may be contested, but the direction is not. A growing share of what scrolls past our children’s eyes synthetic content spun out by machines.

AI tools generate, translate and recombine text, images, audio and video at negligible marginal cost. Studies of synthetic media on platforms like X show spikes in AI‑generated images and videos after each major model release, including viral deepfakes of public figures3 [3]. Misinformation researchers now treat AI‑generated content as a central risk to the integrity of our information environment4 [4].

It is not just about opening a floodgate of information. That flood now also contains:

• more hallucinated facts — confidently wrong answers that sound right,

• more false news and deepfakes, and

• less ability to tell who — or what — actually created what we see.

For a teenager trying to understand the world, signal and noise are becoming harder to distinguish. This is classic information overload, amplified by synthetic media. In response, a cottage industry of “AI detectors” has sprung up. The problem is structural: generators improve continuously; detectors are always one step behind. Europol’s analysis warns that as synthetic media proliferates, technical detection alone will be insufficient; human judgment, contextual verification and more old-fashioned attention will be essential. In other words, the real defence has to live in our minds [2].

That means cultivating a different way of reading and watching:

• Ask high‑quality, grounded questions with enough context. AI systems are pattern‑matchers, not oracles. The more specific your question and situation, the easier it is to see when an answer “sounds right” but clashes with basic facts or lived experience.

• Pre‑empt your own confirmation bias. AI is far too willing to agree and flatter. Before you ask, ask yourself: What evidence would change my mind? Otherwise, you risk using smart tools to dig yourself an even deeper intellectual trench.

• Practice critical, balanced thinking. Check sources, compare perspectives and stay alert to gaslighting, missing context and plausible nonsense dressed up as authority.

These are the cognitive habits that turn AI from a hallucination machine into a thinking aid. They are also habits that children can learn — but only if adults model them.

3. How are we using AI now?

Millions of people now use AI every day. Understanding people’s interactions with AI is one of the great sociological questions of our time. Anthropic, creator of Claude.ai, recently designed a privacy-preserved tool, Anthropic Interviewer, to asks people directly (detailed interviews at unprecedented scale) to get a comprehensive picture of AI’s changing role in people’s lives, including how people are actually using Claude’s output and how do they feel about it. This is a new step in understanding the wants and needs of our users, as well as gathering data for the analysis of AI’s societal and economic impacts5 [5].

Key Usage Trends from Anthropic Interview research results, the Anthropic Economic Index and the AI Fluency Index (late 2025 / early 2026):

• Dominant Uses: Usage is concentrated, with over one-third (36%) of Claude.ai conversations focusing on software development and coding, although educational and scientific tasks are rising.

• Automation vs. Augmentation: While AI agents have spurred an increase in automation (direct task delegation), a significant portion of users still prefer “augmentation”—using AI as a collaborative, interactive, and iterative thought partner.

• Agentic Feature Shift Over Time: By November 2025, 52% of interactions were classified as augmented, while 45% were automated, showing a shift back toward collaboration as more “agentic” (proactive) features were introduced.

• “Artifacts” Impact: When using the “Artifacts” feature (for creating documents, code, or apps), users tend to be less critical, questioning the AI’s reasoning 3.1 percentage points less often than in standard chat, suggesting higher trust in polished-looking outputs.

The trend towards agentic use cases is accelerating. It would be important to take a pause to understand what AI is doing to our brains and what skills would be required to properly leverage AI in amplifying human.

4. What AI is doing to our brains: cognitive offloading and deskilling

There is also a quieter, neurological risk: what happens when we lean on AI too much. And even experts are not immune. But let’s examine the baseline first, as illustrated in a recent MIT Media Lab study led by Nataliya Kosmyna, volunteers wore EEG‑like headsets while writing short essays and taking math tests under three conditions: using only their own brains, using a search engine and using ChatGPT as a co‑pilot6 [6]. The results were telling:

• In the brain‑only condition, participants showed the richest, most distributed brain connectivity, especially in regions linked to attention, planning and memory.

• With search engine assistance, connectivity dropped.

• With ChatGPT co-pilot, connectivity dropped roughly halved on some measures compared with the brain‑only baseline. Participants in the heavy‑AI condition were also unable to remember clearly what they had written later.

A single practice of using AI does not necessarily shrink our brains. But over time, if we rely on AI tools every day for years, we repeatedly offload effortful thinking to AI. In other words, we are not just using a tool, we are training ourselves not to think. That process of cognitive offloading is artificial ignorance in its purest form: high apparent output, low genuine engagement.

The more we lean on AI, the easier it becomes to let judgment idle, even in domains where we are supposed to be the experts. In the Harvard–Boston Consulting Group “jagged technological frontier” experiment, hundreds of BCG consultants were assigned to solve realistic business problems with and without GPT‑4. When they used AI on tasks inside their domain — say, telecom specialists on telecom cases — they completed 12.2% more tasks, 25.1% faster, and with 40% higher quality compared to those not using AI. But when put the same specialists on tasks outside that frontier, performance fell. Error rates rose by 19%, and consultants began relaying AI’s confident but wrong recommendations instead of interrogating them. Over‑reliance turned experts into novices, like a driver falling asleep at the wheel with cruise control on: safe on straight highways, dangerous on sharp bends7 [7].

Deskilling also appears in the very domains where experts are strongest. A study in The Lancet Gastroenterology & Hepatology followed endoscopists after they introduced AI systems to assist with polyp detection during colonoscopy. AI‑assisted procedures improved detection in the moment, but several months later, in unassisted procedures, the doctors’ own adenoma detection rate appeared to fall by about 20%, suggesting a deskilling effect. AI sharpened the tool but dulled the surgeon8 [8].

The remedy is not to abandon AI, but to build in “AI holidays” — regular AI‑free practice that keeps human skills alive even as machines assist [8]. If this is what over‑reliance can do to expert cognition and performance, it is not hard to imagine what happens when still‑forming minds of the children lean on AI for more and more of their thinking. The deepest risk is that children never fully develop the habits of attention and effort that deep thinking requires.

5. AI-amplified attention casino, loneliness, anxiety and mental health exacerbation

Now move from cognition to attention. When AI is implemented in social media and smartphones, it further fragments our focus by supercharging personalised feeds and content generation. Welcome to the attention casino.

Psychologist Professor Angela Duckworth notes a worrying pattern: where students once stayed with a task for around three minutes before switching, the rise of short‑form, highly curated feeds seems to have cut this to well under a minute. The exact “45 seconds” figure is not a law of nature, but the direction is clear: attention slices are getting thinner9 [9].

You cannot build deep expertise — or deep relationships — 45 seconds at a time.

This is not simply about willpower. Research from the University of Portsmouth and the University of Surrey finds that young adults with higher loneliness and anxiety are more prone to problematic smartphone and social‑media use. They often turn to their phones to cope, only to find that compulsive checking and late‑night scrolling make their anxiety worse. AI‑driven recommendation engines sit on top of that vulnerability, optimising for engagement, not well‑being.

Jonathan Haidt, in The Anxious Generation, offers three practices that, uncomfortably, describe many parents’ failures in fighting the attention crisis amplified by AI10 [10]:

• Treat the phone as an experience blocker, not just a distraction. It does not only steal minutes; it can steal entire childhood “sensitive periods” for learning social skills and independence.

• Scaffold real‑world risk. Children do not just need protection; they need difficult projects, physical challenges and unfamiliar groups that build anti‑fragility.

• Fight the algorithm, not the kid. Our children are not weak. They are up against billion‑dollar AI systems tuned to keep them glued to a screen. They do not need more shame; they need allies who understand the game.

The same logic extends into mental health.

On paper, Gen Z is the most connected cohort in history. Yet surveys across countries show rising loneliness and anxiety among teens and young adults. Digital habits are not the only cause, but they have become a powerful amplifier.

AI‑powered companions and “therapist” chatbots plug straight into that vulnerability. Xingye, an AI companion mobile app developed by AI powerhouse MiniMax, has around half a million daily users in China, many of them teen girls and young women11 [11]. Journalist Poppy Koronka reports that children using chatbots from Meta as therapists may see their mental health worsen. US regulators have opened investigations into AI therapy bots over misleading claims and data practices12 [12]. One clinical worry is structural: human therapy is bounded in time and space; sessions end. AI does not have office hours. A child lying awake at 2 a.m. can spend hours ruminating with an endlessly responsive bot with always-on relief, reinforcing anxious loops instead of disrupting them.

It is worth noting that AI can also support healthier habits — for example, by guiding exposure therapy, structuring journalling or offering language practice — when embedded in thoughtful products and bounded routines. But those designs remain the exception.

6. Skills for a human–AI symbiotic balance

Pull these threads together — synthetic content, artificial ignorance, attention slicing, AI‑mediated coping — and one conclusion emerges: skills for a human–AI symbiotic balance sit at the centre of a new parenting playbook. We are not just managing devices; we are shaping the relationship between our children’s minds and an always‑on layer of machine intelligence.

Human–AI co‑intelligence is less a tug‑of‑war and more a sideways infinity loop: one side human, one side machine. At different ages and in different tasks, one loop should swell while the other shrinks — sometimes the child leads and the AI merely suggests; other times the AI drafts and the human edits. The balance is not automatic; it needs deliberate, ongoing calibration.

Three skills matter most.

1. Asking good questions in the right context.
This is the antidote to both hallucination and shallow thinking. It forces us to slow down, frame problems clearly and engage our own cognition before outsourcing the rest. With teens, that might mean insisting they write their own first paragraph before asking an AI to help; with adults, it might mean defining success criteria before letting an AI agent act.

2. Judgment and discernment.
This is the daily practice of verifying claims, cross‑checking sources, resisting easy answers and being willing to update beliefs in light of evidence. AI will keep getting faster and smarter; the question is whether we, as families and communities, can get wiser at least as quickly — or whether we drift down the comforting glide path into artificial ignorance.

For adults and professionals, these two skills translate into clear guardrails. Humans stay in the loop (AI suggestions remain drafts until a responsible person signs off), AI assists but does not replace (co‑pilot, not pilot), and we schedule regular “AI‑off” sessions (or AI holidays) so people practice key skills without autopilot. In high‑stakes domains, that can mean dual‑pass reading, credentialed access to powerful tools and audits of when humans override or rubber‑stamp AI decisions.

3. Human–AI balance as a parenting habit.
For parents, the balance starts with a different set of questions. With teens, it means deciding together where AI should help and where it should stay out: which homework tasks are AI‑assisted versus AI‑free, which creative projects can use AI as a sparring partner versus a ghostwriter, and how much screen time goes to auto‑playing feeds versus deliberate research. You are not banning tools; you are co‑designing the loop.

I find it useful to picture the human–AI symbiotic partnership as an infinity symbol: one loop for the human, one for the AI. For any given task, age or situation, the loops should be different sizes — sometimes the human side dominates and AI only nudges; other times AI handles more routine work while the human decides what matters. But the human loop never disappears; keeping the human in the loop (HITL) is critical. The exact calibration of human and AI roles depends on two skills: asking high‑quality questions with enough context, and exercising judgment and discernment about when to trust, challenge or ignore what the machine suggests.

For younger children, parents can borrow Clayton Christensen’s “Jobs to Be Done” (JTBD) lens13 [13]. Stop asking “Why is my kid using this?” and start asking “What job are they hiring this for?” If a child is using AI “for homework”, is the real job avoiding boredom, chasing quick praise or actually learning the material? Do not fight the tool in the abstract. Ask what job your child is hiring it to do — and whether AI is truly doing that job well for their long‑term growth, or quietly doing the opposite.

Three concrete experiments can make this real in a single month:

• Choose one family activity — a project, trip or meal — that is planned and executed with no AI at all, simply to feel what attention without autopilot is like.

• Have one explicit JTBD conversation with your child about an app or AI tool they love: what job it is doing for them, and whether it is doing that job well.

• Set one clear boundary on AI use for schoolwork (for example, “AI may critique your draft but not write it”) and stick to it.

In the end, ambient intelligence will seep into every corner of our children’s lives. The open question is not whether they will grow up with powerful AI, but whether they will grow up with the inner skills to decide, moment by moment, when to lean on the machine — and when to leave their own minds fully in charge.

Fellow parents, let’s help one another and our children step confidently into the age of artificial intelligence, without sleepwalking into artificial ignorance.

References:

I’m a Gen X dad (born in the 70s) with two teenagers, which means I grew up rewinding cassette tapes and my two kids (born in the 00s) grew up rewinding YouTube videos.

In 2025, that gap hit me harder than ever as my son turned 18 at the end of last year. We spent late nights talking about university choices and future plans, and in May I watched him deliver his valedictorian speech at his high school. Observing him and his Gen Z peers (born in 00s and 10s) forced me to look beyond my own household. Their worldview has been shaped by technology disruptions (smartphones, social media, AI), natural disruptions (pandemics, climate change) and socio-economic disruptions (financial crises, political conflict, wars).

Understanding these generational differences is a survival skill for modern parenting. It changes how we communicate, how we interpret our children’s choices, and how we decide when to step in—or step back.

1. Ready to learn vs ready to teach

Like most parents, I started out believing “good parenting” meant opening as many doors as possible and pushing opportunities through them: extra classes, advice, internships, networks. It was all push.

After 19 years of experimentation (with ongoing debugging), I’ve realised how one‑sided that can be. If parenting were software, I am definitely still in early beta phase. Love may be unconditional; learning isn’t. For the first 16 years, I learned that there is no guaranteed download—no understanding—just because you are broadcasting wisdom on a strong signal.

Then, over the last few years, almost magically, something flipped.

In the middle of Grade 10, my son went from cruise control to full throttle. He began to initiate. He started asking questions, exploring options, thinking seriously about his future. He was no longer just being educated; he was educating himself. He was ready to learn.

By Grade 12, he had a better research stack than I did—search engines, AI copilots, and a willingness to DM/PM the right people on social media. My daughter, two years younger, started her own awakening earlier around Grade 9.

In How Will You Measure Your Life?, innovation theorist Clayton Christensen captures this with a simple but profound insight[1]1:

“Children learn when they’re ready to learn, not when we’re ready to teach. If we’re not with them as they encounter real challenges, we’re not there when they’re finally open to learning. Our job is to be present when life opens a learning window.”

Parenting stops being a broadcast model and becomes more like being on call: the job is to be around when life raises a ticket they care enough to open.

Even this is easier said than done, and at the same time you will also find yourself navigating through a minefield of classic traps along the way, most of which you would only notice in hindsight [1]:

· The helicopter trap: pouring in resources but starving them of experiences that build capability.

· The sequencing trap: “I’ll invest time when I’m less busy.” By the time we’re less busy, their values and priorities are already formed.

Timing and presence turn out to be the scarcest resources, not money or advice. They seem to be the core of effective parenting.

2. Prepare your child for the road, not the road for your child

“One generation plants the trees, and another enjoys the shade” or “前人栽树，后人乘凉” is a beautiful Chinese proverb. It sounds comforting. But in 2025, it can also be dangerous.

Gen X did plant trees. We built careers and companies in a relatively stable globalisation era. But we planted those trees for our climate. The climate has changed—economically, technologically, geopolitically.

If our children simply sit in the shade of what we built, they may be shielded from discomfort, but also from the practice they need to face their own storms. Our job is not to pave the road; it’s to prepare the driver. We don’t get to design the world they inherit; we only get to influence the operating system they bring into that world with.

We cannot teach them our solutions because their problems are structurally different. Christensen reminds us again that our job is not to script their paths but to be the safety net when the world teaches them hard lessons [1].

This also explains another familiar saying: “Wealth does not pass three generations / 富不过三代.” It often doesn’t if the second and third generations only inherit shade from the first, without learning how to plant, prune and replant in different soil. The problem is that shade feels comfortable—right up until the weather changes.

“Prepare your child for the road, not the road for your child. If they only inherit our shade, they won’t learn how to plant in a new climate.”

3. Worldview differences between the operating systems of Gen X and Gen Z

Psychologist Jean Twenge offers a simple way to think about the gap between generations. She argues that technology is the primary driver of generational change: older cohorts lived a “fast life” strategy (grow up fast, marry early, work early), while Gen Z runs a “slow life” strategy in a more complex, high‑stakes world. Seen that way, when Gen Z “quiet quits” or treats work–life balance as non‑negotiable, they’re not slacking; it’s a rational response to a harder game [2]2. Caitlin Fisher calls this “gaslighting”: tell a generation “work hard and anything is possible,” then blame them as entitled when the system breaks under them [3]3.

Generational researchers like Bobby Duffy adds another layer: our “formative years” (roughly age 0–15) are when the outside world quietly writes itself into our mental code. What happens in the society in those years—booms or busts, stability or crisis—shapes the operating system we carry into adulthood [4]4.

For parents, this means our kids’ reactions to work, risk and technology are often about the world they grew up in, not just their individual character. To see how this plays out in my son and me, it helps to look at our formative years and at two speeches we gave in the summer of 2025. I was addressing over 600 graduands at the 214th congregation of the Faculty of Medicine of the University of Hong Kong (available here). My son was delivering his valedictorian address to his cohort Class of 2025 at his high school [5]5 [6]6.

My formative years (70s and 80s): improvable systems

I grew up in analog Hong Kong. Systems did not feel perfect, but they felt improvable. ICAC [7]7 cleaned up corruption. Public housing expanded. Education opened up. Factory kids became office workers. Deng Xiaoping’s “Open Door” economic reforms and globalisation lifted growth. Hong Kong transformed into an international financial centre, and the Sino–British Joint Declaration set the stage for the handover. The message to my generation was simple: work hard, keep your nose clean, and things will generally improve. Systems were not perfect, but they were improvable.

My son’s formative years (10s and 20s): perpetual crises

My son arrived just as the global financial crisis was triggered by the collapse of Lehman Brothers, hardly the soundtrack you would choose for a memorable childhood. It was followed by equally turbulent noise: climate anxiety, deglobalisation, the Umbrella Movement, the 2019 social unrest, COVID, smartphone and AI disruptions. Don’t get me wrong, there were positive events in that period. But by and large, he came of age watching Millennials (generation before him) play by the rules—study hard, get the degree, grind in the job market—and still lose on things like jobs, housing and debt.

Four specific contrasts between the two speeches stand out. It was like watching two different versions of the same operating system.

What is growth to you?

Father: I talked about Felix—my inner alter ego, the voice that challenges and pushes me. I spoke about “manifestly important and nearly impossible” problems and failing forward. My underlying message: the answers are inside you if you’re prepared to wrestle with them.

Son: He described high school as training an AI bot. He compared growth to machine learning: feed in data, make mistakes, adjust the model.

The metaphors are telling. For me, growth is reflecting with an inner self. For my kids, growth is fine‑tuning a model with the outside world.

Definition of resilience

Father: Resilience means endurance and pivot. “Fail fast, fail big, fail forward.” It is about surviving large losses and rebuilding—classic antifragility.

Son: Resilience means adaptation. “Less about powering through and more about learning how to adapt.” It is about learning to bend with disruption (COVID, AI) instead of snapping.

View of AI

Father: AI is a tool. I challenged graduates to use AI for the obvious tasks and reserve their humanity for what AI cannot do: compassion, ethics, complex judgement.

Son: AI is a companion and a model. He joked about ChatGPT as his “new buddy” and used AI itself as the metaphor for human learning. For him, AI is not just a tool in the environment; it is part of the environment. He is near-AI-native.

Mindset

Father: The answers are inside you, if you have the grit to dig them out.

Son: The answers emerge from interaction—between self, peers, systems and algorithms.

These contrasts are not just differences in speechwriting; they are signs of a deeper rewiring in how each generation understands risk, time and growth. If technology could shift my son’s mental “operating system” this far from mine, it raises an obvious question: what happens to the children who grow up not just with phones in their pockets, but with AI woven into every corner of their childhood? What does the next great rewiring look like?

4. The next great rewiring, this time by AI & Robotics – Hello Generation AIR and Q

Social psychologist Jonathan Haidt calls 2010–2015 the “Great Rewiring” of childhood: the shift from play‑based to phone‑based childhood. Unstructured outdoor play was replaced by structured activities and always‑on screens, contributing to anxiety, sleep loss and social fragility in Gen Z [8]8.

But next great rewiring, even more powerful than phones and social media, is already underway: AI and robotics. Generation Alpha (born roughly from 2013 to the mid-2020s), the children coming after my kids, are already feeling its effects, but the cohort after them, Generation Beta (born from the mid-2020s onwards) will be thoroughly shaped by AI and robotics. If we were to choose a more directly descriptive name for Generation Beta, it could be called Generation AIR (Gen AIR): AI & Robotics natives, for roughly those born from 2020 to 2045. After Gen AIR lies Generation Q (Gen Q), who may grow up with quantum computing and bio‑digital hybrids as normal. At that point, our nostalgic arguments about radio vs television will sound very quaint.

If Gen Z grew up with phones in their pockets, Gen AIR will grow up with AI in their bedrooms, classrooms, toys and all devices. Recent report in The Economist describes AI tutors that adapt perfectly to a child’s level, toys that talk back and remember their preferences, and AI companions that are always available, never sulk, never argue and always validate their feelings. That “perfect companion” creates new kinds of risk: echo chambers, “yes-bots” and cognitive offloading [9]9.

Echo chambers for children: the football‑obsessed child gets only football stories, examples and games. Imagine a childhood where the algorithm decides you’re “the football kid” at age six and never lets you forget it. Serendipity vanishes; so does tolerance for the unfamiliar. This is similar to the social‑media algorithm that feeds you content solely based on your preferences and exacerbates echo chambers and confirmation bias.

“Yes‑bots” as friends: a third of American teenagers (and likely the same in other developed countries) already say talking to an AI companion feels at least as satisfying as talking to a friend, and easier than talking to parents. For a teenager, a friend who never disagrees can be very tempting—and very misleading training for real relationships. My teenage self would have loved a chatbot that did my homework. My adult self is less sure.

Cognitive offloading / Cognitive delegation: students using AI tools show less brain activity and recall less of their own work, suggesting they are outsourcing not just tasks but thinking itself. A recent MIT study measured brain activity with EEG while three groups of people performed tasks using only their brains, a search engine, or an AI/LLM respectively. The results showed clear differences in neural engagement [10]10:

· Users relying solely on their brains had the richest, most distributed neural activity.

· Search engine users had reduced neural activity.

· LLM users had the lowest. Many users could not recall what they had written in their essays.

Simply put, when we lean on AI too heavily, our brains idle. Over time, this erodes learning capacity — a path straight to what I call “artificial ignorance.” For the trained professionals, “artificial ignorance” also means deskilling of their mastery.

If Gen Z is “anxious”, according to Haidt, Generation AIR risks being completely hollowed-out, raised with “perfect partners” who never demand compromise, patience or negotiation, yet quietly encourage subconscious cognitive offloading.

5. A bridging toolkit: Jobs to Be Done

Standing here as a Gen X parent with Gen Z kids, watching Gen AIR on the horizon, I find myself reaching back to Clayton Christensen for help. His “Jobs to Be Done” framework suggests we stop asking “Why is my kid using this?” and start asking “What job are they hiring this for?”[1]

When my child is glued to a screen, that device is doing a job:

· Functional: entertain me, teach me, help me finish homework faster.

· Social: keep me in the group chat, let me belong.

· Emotional: make me feel competent, comfort me, numb my anxiety.

In my childhood, the job “help me feel competent and independent” was served by roaming the neighbourhood, taking buses alone, finding part‑time work. In my son’s childhood, the same job might be served by mastering a game, building an online project or creating content.

The human need or the job is the same; the hired solution looks different. Instead of fighting the tool (screen, phone, social media, AI, etc.), we can ask what job it’s doing, check if it’s doing that job well, and offer alternative ways to meet the same need in the physical world.

“Don’t fight the tool. Ask what job your child is hiring it to do—and whether it’s actually doing that job well.”

Gen X and Gen Z are in a unique position because they are the bridge. Gen X remembers life before digital, mobile phone and the internet; Gen Z remembers life before AI, Robots and Drones. Together, they bridge analog, digital and now AI‑native childhoods.

My kids’ generation (Gen Z) will parent, teach and manage Gen AIR and Q. They have to figure out how to raise humans in a world where AI and quantum computing can do almost everything better, faster and cheaper—hopefully except be human.

6. An experimental prescription for the Human OS

When I read Jonathan Haidt’s three simple guidelines in Anxious Generation, I recognize some of my own mistakes. I’ve pinned these guidelines to my own mental dashboard [8]:

Treat the phone as an experience blocker, not just a distraction. It can block critical “sensitive periods” for learning social skills. Our job isn’t just to confiscate or shame the child; it’s to reopen those periods: family dinners, walks, shared projects, awkward but real conversations.

Scaffold real‑world risk. Encourage the kinds of risk we grew up with—travel, hard projects, unfamiliar groups—so they can rebuild some of the antifragility that unsupervised play once provided.

Fight the algorithm, not the kid. Our children are up against billion‑dollar recommendation engines. They don’t need another enemy; they need an ally who understands the game.

Which brings me back to our two speeches.

In my HKU talk, I ended with a health‑life prescription such as: “May you detect the first cancer cell of ego and heal them daily with kindness and humility.[5]”

In my son’s valedictorian speech, he told his classmates that in the long run, their lives would be measured not by grades but by “the relationships we’ve built, the kindness we’ve shown, and the lives we’ve touched.[6]”

Different worldviews but the same GPS coordinate: kindness. Perhaps this is our Human Operating System (Human OS).

Everything else is upgradable or replaceable: devices, platforms, languages, job skills. The Human OS—kindness, resilience, curiosity, courage, deep relationships—is the only software that has remained backward‑compatible through every generation.

“Our parenting playbook is obsolete. Our Human OS (kindness, resilience, real connection) isn’t.”

We don’t get to captain our children’s ships anymore; they have better navigation tech than we do. But we can be lighthouses—steady, visible, honest about the rocks and the storms.

That, at least, is the faith this Gen X dad is choosing to have. A century ago, people feared that the diffusion of colour TV would make society shallow and destroy the Human OS. The same fears appeared with computers, the internet, social media and smartphones. AI and, later, quantum technologies will stress it in new ways, but if history is any guide, the core code of the Human OS will adapt and survive. And if my kids are reading this: yes, I still reserve the right to complain about your screen time, lovingly so.

Wish you an awesome 2026!

Notes: I have used Perplexity to perform final edits and Gemini 3/Nano Bananas for the last illustration.

Footnote references: