Technology: Phones, Social Media, AI Tools | Raising Humans for an Unknowable World Tutorial

The Current Situation

A child born in 2020 will grow up with technologies that didn't exist, or were minor, when their parents were children. Smartphones, ubiquitous internet, algorithmic social media, AI tools. These aren't features of daily life; they're the environment.

The parenting challenge: how do you shape a child's relationship with these technologies, at various ages, when the technologies themselves are changing faster than the advice about them?

This chapter takes a position. It tries to be honest about where the evidence is strong and where it's uncertain. The position leans cautious; your household will land where it lands.

What the Evidence Suggests

The most thoroughly argued case against a common current default: Jonathan Haidt's work, summarised in The Anxious Generation (2024). His claims, compressed:

Phone-based childhood hurts kids. Smartphones plus social media, in children under ~16, correlate with rising rates of anxiety, depression, and self-harm, especially in girls
The changes in mental health are too large and too coincident with phone adoption to be explained by other factors alone
The mechanism involves social comparison, sleep disruption, sedentariness, reduced face-to-face contact, and exposure to harmful content
The fix is collective: delay phones until high school, social media until 16, phone-free schools, and a return to more unstructured in-person play

Haidt's case is contested (some researchers find his evidence less decisive than he does), but directionally it's consistent with a lot of what parents observe and what the data shows.

What you can probably bank on

Compiling across the evidence:

Smartphones in the hands of young children are bad, on average, for development of attention, independent play, and face-to-face social skills
Social media algorithms optimised for engagement are not optimised for wellbeing
Sleep disrupted by devices is meaningful and measurable
Screen time at the expense of physical activity and social interaction has costs
The platforms' age policies (usually 13+) are not evidence of safety; they're compliance markers

None of these are controversial enough to wait on. You can act on them now.

What's genuinely uncertain

The long-term psychological effects of AI companions (Replika, Character.AI, future ones) on children and teens
How AI tools affect learning and skill development (net positive for some skills, net negative for others, and the mix isn't settled)
The specific age at which different content risks matter most
Whether the harms generalise across cultures or are specific to anglophone contexts studied most
What balance of restriction and engagement teaches the best long-term relationship with tech

The uncertainties don't paralyse action; they suggest humility about specific rules.

The Core Problem

The core problem isn't "technology is bad". It's that many current technologies are designed to be maximally engaging. Engagement is what they optimise for, because engagement is what produces advertising revenue.

Maximally engaging is not the same as beneficial. Slot machines are maximally engaging. So are cocaine and sugar. So, apparently, are algorithmic feeds.

A child given access to maximally-engaging technology without supports tends to use it in ways that feel good in the short term and don't serve them long term. This is not a moral failing of the child; it's a predictable outcome of the design.

Parents are, in effect, pushing back against well-resourced optimisation. The platforms have more money and more behavioural scientists than you have time and attention. The asymmetry is real.

Principles Rather Than Rules

The specifics of what to allow change as technologies change. Principles last longer:

1. Delay

Wait longer than the default before introducing new technologies. When smartphones became available to 10-year-olds, "delay" meant 13. When they became standard at 10, "delay" means 14 or 15. The target keeps moving.

The principle: there's no urgency to introduce these. A child without a smartphone at 10 is not deprived; they're having a childhood.

2. Unstructured first

Make sure the child has substantial unstructured, device-free time before devices enter. Outdoor play, boredom, reading, imaginative play. These are the raw materials of development.

If devices fill the available time, the raw materials are absent.

3. Restrict what you can't supervise

As children age into using devices, restrict categories you can't usefully oversee. Adult content. Algorithmic social media that optimises engagement over wellbeing. AI chat that can form parasocial relationships.

Being the "cool parent" who allows everything is not a goal worth having. You can be warm and close while still restricting clearly.

4. Model use that resembles what you want them to have

Your own phone relationship is a teacher. If you're on your phone constantly, looking at social media, distracted during meals, they learn that this is adult life. If you have a more limited relationship, they see that option.

This is one of the hardest specific parts of modern parenting. Most adults have their own unhealthy relationship with phones. Working on your own is part of the job.

5. Connect through tech, don't delegate to it

Tech used together is different from tech used alone. Watching a film with your teenager and discussing it beats each of you consuming separate screens. A shared video game can be connective; a feed each scrolls separately is not.

6. Phones off in key places

Bedrooms at night. Meals. Homework time. Whatever your family's rules are, make them consistent and apply them to yourself.

The specific rules matter less than the principle: there are spaces in family life that phones don't colonise.

By Age, Roughly

Rough guidance; adapt to your child and context:

0-5

Minimal screens; ideally video calls with family only
If videos: short, educational, co-watched. Not standalone entertainment time
Music and audiobooks are different from video and generally fine
Devices out of the bedroom entirely

6-10

Limited screen time; clear daily limits
No personal smartphone
A family tablet with curated apps, used in shared spaces
Video games (if at all): short sessions, age-appropriate, co-played or at least observable
No social media

11-13

Still no smartphone if you can hold the line
A "dumb phone" (calls and texts only) is a viable intermediate
Limited social media if any; no algorithmic feeds
Clear rules about screens and sleep
Starts of conversations about online life

14-16

Possibly a smartphone, with real restrictions (no algorithmic social media, screen-time limits, supervision settings)
Conversations about AI, pornography, misinformation, online predation (yes, you have to have these)
Modelling how to use tech as a tool rather than succumbing to it
Increasing autonomy paired with continued visibility

17-18

Gradual transition to full adult use, with expectation they can manage it
By 18, they should be capable of self-regulating, or you've missed the window

These are provisional. The right answer for your family depends on their temperament, your community's norms, and your judgment.

The hardest part: you're coordinating against local defaults. "All her friends have phones" is a real pressure, and not entirely wrong. A child who is the only one without the technology that mediates their peer group is isolated.

Partial solutions:

Find families with similar values and coordinate delays. One child alone is isolated; five children aligned is a subculture
Accept some isolation, especially early. Better to be slightly outside the peer group now than to damage long-term development
Schools matter: phone-free schools remove a lot of pressure. If you can help your child's school move in that direction, the effort pays back
Talk about it: your child should know why. "You'll get a phone later than your friends because I think the evidence suggests it's better for you" beats "because I said so"

AI Specifically

AI tools are newer and their effects on children are less studied. Some thoughts, tentative:

Using AI as a tool

Older children (teens) will increasingly use AI for research, writing, coding, other tasks. This is part of their future; pretending otherwise is a disservice.

The goal: use AI as a tool to develop their own thinking, not a replacement for it. A child who uses AI to draft an essay and then doesn't learn to write is worse off than one who writes their own drafts and uses AI to get unstuck.

Parents can't police this directly. They can shape the frame:

Discuss what AI does well and what it doesn't
Distinguish "AI helped me think" from "AI did it for me"
Value the skill of writing/coding/thinking independently of whether AI can also do it
Notice when work seems AI-generated and have conversations about it

AI companions

Apps that present AI characters for conversation and relationship are a real concern for children. The parasocial attachment can be strong; the effects on social development are not yet studied.

A reasonable default for young teens: no AI companions. For older teens: awareness and conversation rather than prohibition.

Deepfakes and misinformation

A specific literacy skill: being able to treat images, video, and audio as potentially fabricated. Teaching basic scepticism is now a parenting task.

Show them examples. Discuss how to verify. Normalise checking sources. This is part of adapting children to the information environment they'll actually encounter.

Your Own Relationship with Tech

An uncomfortable truth: you can't produce a healthy child tech relationship if yours is broken.

Honest questions:

How many hours a day are you on your phone?
Do you check it during conversations, meals, bedtime?
Do you use social media algorithmically (doom-scrolling) or deliberately?
How often are you actually absent while physically present?

If the answers are worrying, address them. Not perfectly; perfection isn't the target. But visibly working on them.

Common Pitfalls

"I'll trust my child to self-regulate." Self-regulation is developed over years, against systems designed to overcome it. Expecting a 12-year-old to self-regulate is expecting them to win a rigged game

"Every parent is doing this." Many are. Some aren't. Coordinate with the ones who aren't. Accept you may be slightly out of step

"Technology is inevitable; I'll just let them have it." Much about how technology is introduced is in your control. The specific technologies are inevitable; the timing and manner are not

"They need to be digital natives." They will be, whatever you do. The advantage isn't in earlier adoption; it's in healthier relationship to the tools

"My parents were strict; it made me worse." Plausibly true for your generation. Specifics don't transfer. What the evidence says now is what matters now, and the evidence isn't kind to unfettered access

Next Steps

Continue to 10-the-hard-tradeoffs.md for the moments when invariants pull against each other.