When a child is bullied online, exposed to harmful content, or drawn into dangerous spaces, our instinct is often to ask: how do we stop children encountering harm?
It is an understandable question.
But I sometimes wonder whether we are asking it because it feels easier than asking a more uncomfortable one: why is so much harm being created in the first place?
If a river is carrying bodies downstream, eventually we have to look upstream.
⸻
When people say “social media is toxic,” what they usually mean is that human beings are using social media in toxic ways.
When people say “AI is dangerous,” what they often mean is that human beings may use AI dangerously.
The distinction matters, because it tells us where to look.
We have become very good at teaching children how to navigate danger. Internet safety lessons. Privacy settings. Parental controls. Age verification. Restrictions.
Yet we seem far less interested in understanding why so many people are creating the very dangers we are trying to protect children from.
Bullying does not begin with technology. Cruelty does not begin with technology. Prejudice does not begin with technology.
Technology amplifies what already exists. The internet did not invent exclusion, humiliation, scapegoating or dehumanisation. Human beings did.
The same technology that exposes one child to harm may allow another to find friendship, belonging and community. The same AI tool that can spread misinformation can also help a struggling student learn. The same gaming platform that worries policymakers may be the place where an isolated autistic child has their closest friendships.
The question is not whether technology is good or bad. The question is what kind of society we are building around it.
⸻
If children are encountering increasing hostility online, perhaps we should also be asking what they are seeing modelled offline.
Consider how the public conversation has shifted around disability. The growing talk of overdiagnosis, of an ever-rising benefits bill, of who is really struggling and who is faking it. None of this stays in policy documents. It drips down into how people think, and then into what they feel entitled to say. Before long it feels acceptable to treat disabled people as a drain, as suspect, as less than.
The same pattern plays out elsewhere. The framing of refugees and migrants in the media and by public figures starts tentative and picks up pace, until people feel confident saying things online they would once have kept to themselves.
The human brain is already wired to find difference frightening. When we amplify that fear, normalise it, and reward it, it is no wonder that in online spaces it grows and takes on a life of its own.
And it does not stay online.
The permission spreads. What begins as rhetoric in public debate becomes conversation in communities. What begins as a joke becomes an assumption. What begins as a comment online becomes something people feel comfortable saying aloud.
Before long, the family who are openly gay, disabled, neurodivergent, from another country, another faith, or simply different in some visible way discover that the conversation has moved beyond the screen.
And then we wonder why our children copy it.
⸻
Children learn from culture. They learn from us.
What is one person’s flag in the garden becomes another person’s permission. What grows online seeps into the world we live in — and what we model in the world seeps back online, with the brakes off.
We spend enormous energy managing the consequences of problems while neglecting the conditions that created them.
Perhaps protecting children requires more than restricting access to harmful spaces. Perhaps it also requires looking upstream — at who is putting the bodies in the water in the first place, and at why so many people are willing to jump in, and why the river is filling faster than we can pull the bodies out.