The Children Can't Wait
My inbox looks different lately.
It used to be a mix of project updates, grant deadlines, and the usual noise that comes with running training & technology programs for a national nonprofit. But over the past several months, something has shifted.
The requests coming into Zero Abuse Project have changed in both volume and urgency. School counselors. Child protective service workers. Pediatric nurses. Youth-serving program directors. All asking some version of the same question: Can you help us understand what's happening with AI and our kids?
They're not asking out of curiosity. They're asking because they're scared. Because they're sitting across from children who are experiencing things that didn't exist five years ago, and they don't have the language or the framework to respond.
Sextortion. AI-generated abuse imagery. Deepfakes of real kids. Their staff is overwhelmed. Their training materials are outdated before they're even printed.
And this week, the United Nations confirmed what those frontline workers already know in their bones.
When the World Finally Pays Attention
On July 6th in Geneva, the UN convened its inaugural Global Dialogue on AI Governance alongside the AI for Good Summit. Secretary-General António Guterres stood at that podium and said plainly that AI is "developing faster than anyone can keep up."
Here's what came out of Geneva that we all need to sit with. The UN's child safety pledge laid out specific expectations for AI developers:
- Any system accessible to children must undergo independent, child-specific safety evaluation before deployment.
- Zero tolerance for sexual abuse content, with active detection, reporting, and removal built in.
- And this one stopped me completely: systems must be able to recognize a child in distress and route them to real human support.
That last requirement is not a technical specification. It's a values statement. It's saying that when a child reaches out in a moment of crisis, the answer cannot be another algorithm. It has to be a person.
That is trauma-informed design translated into policy language. And seeing it come from a global governance body is something I genuinely did not expect to witness this soon.
For those of us who have been building with these principles quietly, inside nonprofits and mission-driven organizations, for years, this moment feels significant. The world is starting to speak our language.
The Gap Between Policy and the People Who Need It Most
Here's my honest concern, though.
Global pledges are important. International frameworks matter. But the school counselor in rural Missouri, the CPS worker in South Los Angeles, the pediatric nurse in a small clinic in Tanzania, **none of them will read a UN policy brief. **
They will show up to work tomorrow with the same questions they had yesterday and no new tools to answer them.
That is the gap, folks.
The training requests we're fielding at Zero Abuse Project right now reflect something real and urgent. Frontline professionals who work directly with children are in desperate need of practical, accessible education on AI risks, sextortion, image-based abuse, and how to talk to kids and families about all of it. They need it now. Not in a two-year curriculum development cycle. Not after the next federal funding round clears. Now.
And they need it in plain language. Without jargon. Built for people who are already stretched thin and doing some of the hardest work that exists.
What "Child Safety by Design" Has to Mean in Practice
The Geneva pledge essentially makes child-safety-by-design a global baseline expectation. I think that's right. And I think it has implications far beyond the tech companies developing AI products.
It means every organization that deploys AI tools in a youth-serving context carries responsibility for how those tools were built and what safeguards exist.
It means nonprofits can no longer treat AI adoption as purely a productivity question. It means the people training frontline staff need to understand these risks well enough to teach them.
And it's why I keep saying that ethical AI implementation in the nonprofit sector isn't optional anymore. It's part of our duty of care to the communities we serve.
A Note to the Frontline Workers Reading This
If you work with children and you are trying to figure out what AI means for your practice, I want you to know that your urgency is valid. The overwhelm you're feeling is not a personal failing. You are navigating something genuinely new and genuinely fast-moving, and the systems that should be supporting you haven't caught up yet.
You are not behind. You are ahead of most institutions, because at least you're asking the question.
As Guterres said in Geneva, AI is developing faster than anyone can keep up. But that doesn't mean we stop trying. It means we build better bridges between what the technology is doing and what the people doing this work actually need.
The children we serve don't have time for us to figure this out slowly.
I'm going to keep building those bridges, in the training programs, in the tools, in this newsletter, and in every conversation I can get into. I hope you'll keep building them too.
And if your organization needs support figuring out where to start, I'm not hard to find. I, along with my other colleagues, offer training on these topics through Zero Abuse Project .
Check out my training topics here.
Stephanie Hall is the Technology & Innovation Lead at Zero Abuse Project and founder of Lilac Creative. She writes Mission/Shift for the people at the intersection of mission, technology, and doing right by the humans we serve.
