Imagine a room full of the world's most powerful regulators, politicians, and tech executives gathered in a grand Danish palace. The official mission: keep children safe in the age of AI. The actual guest list? A who's who of people who have spent years pushing governments to decide what the rest of us are allowed to read, watch, and say online. That was the scene at the Youth AI Safety Institute summit in Copenhagen, and it should worry anyone who values open conversation and free thought.
Launched just days ago by Common Sense Media, the new "Youth AI Safety Institute" promises independent ratings and standards for AI tools used by young people. Sounds reasonable on paper. But look closer, and the summit reveals a familiar pattern: "safety" framed so broadly that it justifies expanded content moderation, age verification, algorithmic tweaks, and regulatory overreach that doesn't stop at kids — it ripples out to everyone.
Every major voice on stage has a history of advocating for tighter speech rules:
Hillary Clinton has repeatedly called for more oversight of social media and now warns that AI demands "urgent investment" and "accountability" from institutions willing to be "unrelenting."
Ursula von der Leyen, architect of the EU's Digital Services Act, pushes harmonised European rules so platforms, not parents, get further tools to shape what children (and adults) encounter.
Vivek Murthy, former U.S. Surgeon General, championed digital ID systems to fight "misinformation" and worked with tech companies to suppress content the government disliked.
Melanie Dawes of UK Ofcom enforces the Online Safety Act, which already pressures platforms to remove "harmful" content under threat of massive fines.
Others include advocates for weakening end-to-end encryption so private messages can be scanned.
These aren't neutral child-protection advocates. They represent the same coalition that turned "protect the kids" into justification for the EU's DSA, the UK's Online Safety Act, and similar regimes, laws that have already led to over-removal of legitimate speech, pressure on platforms to act as censors, and governments ordering deletions of disfavoured viewpoints.
The Institute explicitly aims to "complement" and implement frameworks like the EU AI Act, DSA, and UK Online Safety Act. Those laws don't just target explicit harm, they create vague categories like "disinformation," "hate," or "harmful content" that give regulators sweeping power. Now they want the same logic baked into AI: pre-emptive filtering, safety ratings that could sideline models allowing open discussion, and "child-centred governance" that inevitably leaks into adult tools.
Parents rightly worry about AI chatbots giving dangerous advice, generating explicit images, or amplifying harmful trends. No serious person dismisses those risks. But the solution isn't handing unaccountable bodies, often funded by the very AI companies they'll rate, the power to define "safety" for the entire internet.
When "safety" means:
Mandatory age verification that requires ID for basic AI use,
Algorithmic suppression of controversial but legal ideas,
Pressure on developers to neuter models so they refuse certain questions,
We're not protecting children. We're building the infrastructure for speech control at scale. Once the filters and ratings systems exist for kids, expanding them to everyone is a small step. History shows mission creep is inevitable.
The Better Path: Responsibility Without Centralized Control
Real child safety doesn't require global regulatory cartels. Parents already have powerful tools: device-level controls, open-source models they can audit, transparent company policies, and most importantly their own involvement in raising kids. Competition among AI providers, clear liability for actual illegal content (not vague "harm"), and technological solutions like better parental filters beat top-down mandates.
The free speech case is straightforward:
Truth emerges through open debate, not curated "safe" outputs.
Innovation thrives when developers aren't constantly looking over their shoulder for the next regulator.
Adults have the right to encounter ideas governments or NGOs deem risky.
Past "misinformation" panics (COVID origins, election integrity, gender medicine) showed how quickly official "safety" becomes enforced error, or terror!
The Copenhagen summit isn't an isolated event. It's part of a pattern where every new technology, social media, then AI, triggers the same reflex: more control, more surveillance, less trust in individuals. "Think of the children" has become the evergreen excuse for eroding liberties.
We can protect kids without sacrificing the open internet that has driven unprecedented knowledge, creativity, and accountability. Families deserve tools that empower them, not bureaucracies that decide for them. Developers need room to build truthful, capable systems, not lobotomised ones afraid of controversy.
The real danger to the next generation isn't unfiltered AI. It's raising them in a world where the default is censored, monitored, and "safe" at the cost of curiosity, resilience, and freedom. Copenhagen shows the elites' vision. The alternative is simpler and more human: trust parents, protect actual rights, and let truth compete in the open., as J. S, Mill once argued in On Liberty (1859).