Full disclosure: I’m a partner in ChildSafe.dev — but my views on child safety didn’t come from the company. I joined the company because of my views on child safety.

For years, digital child safety has meant monitoring. Screen-time dashboards. Weekly reports. Notifications after something went wrong. Parents were given more information, but systems largely stayed the same. That is finally starting to shift.

A growing body of policy analysis — including the recent report from the American Enterprise Institute — points to the same conclusion: we’re moving from parental monitoring to embedded, environment-level protection. AEI makes a critical observation that many miss: child safety works only when responsibility sits in the layer that actually shapes a child’s experience. That aligns directly with the architectural changes unfolding across major platforms.

From tools to systems

If you step back and connect the dots between AEI’s analysis and recent platform updates, a pattern emerges. Operating systems are beginning to pass age signals in privacy-preserving ways, so apps can adjust experiences automatically. Platforms are building “Teen Accounts” with non-negotiable defaults. Game environments are adding experience-level blocking rather than broad age buckets.

These changes all point toward the same truth: child safety cannot rely on perfect behavior — from kids, from parents, or from platforms. It has to be baked into the design. This is progress. But it is not yet a system.

Fragmentation is not an ecosystem

AEI’s warning is blunt: placing safety responsibility in the wrong layer creates a false sense of security. And that’s exactly what we see today.

Parents still manage safety through a patchwork of dashboards, app settings, OS menus, and platform controls. Protections differ depending on device, app, and context. Safety features exist, but they do not travel with the child. That’s the difference between features and infrastructure.

A real ecosystem is coherent. Predictable. Durable. It works even when individual actors fail. Today’s landscape is none of those things — not yet.

Where responsibility actually belongs

AEI’s argument is simple and correct: children don’t experience the internet as one place. They experience applications, each with its own incentives and risks.

1. A messaging app
2. A short-form video platform
3. A game
4. A learning tool
5. A live chat feature
6. A recommendation feed

These aren’t interchangeable. They don’t carry the same risks. They don’t influence behavior in the same ways. So responsibility shouldn’t be centralized in operating systems or app stores — layers that distribute software but do not shape experience. AEI is explicit: that approach is administratively clean but strategically wrong.

Responsibility needs to sit where design choices are made. That’s where risk originates. That’s where risk has to be controlled.

This is a familiar moment

We’ve lived through this transition in other sectors. Safety used to be reactive — warnings, disclaimers, and after-the-fact corrections. Then we recognized the limits of human behavior. Seat belts accepted that people make mistakes. Airbags accepted that sometimes they don’t even use the seat belt. This wasn’t moralizing. It was engineering. Systems scale faster than judgment. When the cost of failure is high, protection must be built in. Digital child safety is standing at that same threshold.

What a real ecosystem must include

A functioning child-safety ecosystem requires alignment across several layers:

1. Education and literacy
2. OS-level signals and defaults
3. Platform-level constraints
4. Developer-level safety architecture
5. Clear enforcement and accountability
6. Legal frameworks that reflect foreseeability and preventability
7. Research the updates and (re)informs the system

AEI is right to warn that putting responsibility at the wrong level undermines the entire effort. If responsibility does not follow design, the system will not hold.

The direction is clear

The past year tells us the direction of travel has changed. Safety is moving:

1. closer to the product
2. closer to the design
3. closer to the architecture

The question is no longer whether safety belongs there. It’s whether we will finish the job. AEI’s analysis is a reminder that shortcuts — especially centralized ones — will not get us where we need to go. Precision and responsibility placement matter. We are closer to a real ecosystem than people realize. The next step is making sure it’s one that actually works — across layers, across platforms, and across the environments where children live their digital lives.