This week made one thing very clear: kids are not waiting for adults to catch up. They are using AI for homework, to ask questions, to figure out things about their bodies and feelings, and sometimes just because they want something to talk to.

And the headlines showed both sides of that. Some kids were at the White House being celebrated for building cool things with AI, which is amazing. At the same time, researchers, school boards, and governments around the world were raising real concerns about safety, privacy, and what happens when kids use these tools without enough guidance.

AI can help kids create, learn, and build. It can also put them in situations they are not ready to handle alone.

If you missed last week’s podcast episode, you can check it out here.

Okay, let’s get into what happened this week.

The White House recognized six student teams as national champions in the first Presidential AI Challenge. More than 20,000 students from all 50 states, D.C., Puerto Rico, and Department of Defense schools in 10 countries took part, producing over 2,500 projects. Winning projects ranged from homework help to bullying prevention to a tool to help visually impaired people navigate. Some education experts, while praising the work, questioned whether students were taught enough about the ethics, privacy, and accuracy risks behind the tools they built.

What parents and teachers need to know: This is the hopeful side of AI literacy. Kids aren't just consuming these tools, they're building with them to solve problems they care about. Then comes the next question, the one the experts are right to ask: a child shouldn't only learn to make an AI project, but to ask "What problem am I solving, who could this harm, and who should it help?" If your student wants to enter next year, start there.

UNICEF launched a policy brief titled When AI becomes a Friend, focused on the risks of chatbots and AI companions that children increasingly turn to for advice, support, and even relationships. The brief argues that conversational and relational AI poses distinct and heightened risks for children, and calls for shifting from reacting after harm happens to preventing it, with responsibility shared across governments, companies, parents, and educators rather than placed on the child alone.

What parents and teachers need to know: Teach kids: "Friendly does not always mean trustworthy." An AI can be warm, endlessly patient, and available at 2 a.m. in a way a tired grown-up can't always be, and that's exactly why a lonely kid can lean on it hard. Name it: it's built to keep you talking, it doesn't actually know you, and it can be confidently wrong.

Two governments acted within days of each other. The UK signaled a sweeping "Australia plus" plan that, alongside a social media ban for under-16s, would require AI "romantic companion" chatbots designed to simulate sexual relationships or roleplay to enforce a minimum age of 18, with similar intimate features restricted for under-18s more broadly. Days earlier, on June 10, Canada introduced a digital safety bill that would ban social media for under-16s and set up a new regulator to make AI chatbots safer, with penalties up to 3% of global revenue. The Canadian bill, introduced after a school shooting, would require chatbots to reduce the risk of harmful content and include crisis-intervention steps, though legal experts have raised concerns about loopholes and its long timeline.

What parents and teachers need to know: It's good to see the "village" stepping up across borders with the same worry: chatbots that build one-on-one emotional relationships with kids. Even with new laws coming, the first protection is still a conversation. Tell your kids, simply and without fear, that if anything they're talking to ever steers somewhere scary, they can come to you, any time, no trouble. I know I keep saying this, but this is what has to happen when there are no full-stop protections in place.

This week, New York's bill to restrict AI "companion" chatbots for minors became national news as it landed on the governor's desk. The bill, S 9051, would bar AI companies from offering companion chatbots to users under 18 and let the state attorney general fine violators up to $25,000 per violation. It won final passage on unanimous votes of 137–0 in the Assembly and 60–0 in the Senate. The governor has until December 31 to sign it, veto it, or let it become law.

What parents and teachers need to know: A unanimous vote on anything is rare, and it tells you how seriously this risk is being taken across the aisle. But a law that may not take effect until next year doesn't protect your kid today. Treat policy as the sloooow safety net underneath you, not the thing standing between your child and a chatbot today.

A story from The 74 and Honolulu Civil Beat looked at how students across Hawaiʻi, from young kids to high schoolers, are experiencing AI in school. Civil Beat spoke with nearly two dozen educators, administrators, and students about their experiences and what they expect the future to look like. The takeaway: AI isn't a far-off thing for these students, it's already shaping how they learn and plan. One McKinley High junior put it bluntly, calling AI "a useful and dangerous tool" that people should be cautious with.

What parents and teachers need to know: Kids are already forming opinions about AI based on what they see at school, online, and from each other. So the question is not “Should we talk to them about AI?” I believe the question is, “Are we helping them think clearly about it?”

In a piece for The 74, DonorsChoose CEO Alix Guerrier shared what he sees across the 90% of U.S. public schools his platform reaches. Since the 2022–23 school year, teacher requests for AI-related tools have surged more than 200%, but the more striking part is the purpose. Teachers are overwhelmingly asking for AI tools to reach students who've been underserved for decades: kids living with disabilities and English learners. One example: a $90 AI translation pen that scans text and reads it aloud or translates it into more than 100 languages.

What parents and teachers need to know: This is AI at its best, opening a door for a child who was being left outside it. Hold onto one line, though: the tool should support the child, not replace the teacher, the aide, or the relationship that helps a kid feel seen. A translation pen is a bridge to a person but not a substitute for one.

At its WWDC conference, Apple announced a new Siri AI, expanded Apple Intelligence features, and a significant overhaul of its child-safety controls. New tools include "Time Allowance," which lets parents set app-category restrictions and schedules, and "Ask to Browse," which lets kids request website approvals through Messages. Apple is also opening frameworks to outside app makers, including a feature to require parental approval before kids connect with new contacts, and a way to tailor experiences to a child's age range without collecting an exact birth date.

What parents and teachers need to know: Better controls are genuinely useful. Turn the guardrails on, then explain the "why" out loud: this isn't because I don't trust you, it's because this thing is built to keep you hooked. Kids respect a boundary far more when they understand what it's protecting. Set it, then explain why.

Common Sense Media released a new survey of more than 1,200 kids ages 9 to 17, one of the first major efforts from its new Youth AI Safety Institute. It found that 86% of kids have used generative AI tools like ChatGPT and Gemini, and nearly a quarter use AI every day. The report looked at how, why, and how often kids use AI, and flagged concerns about both educational and emotional dependency. CEO Jim Steyer framed the findings as a "pretty big wake-up call" for parents, educators, and policymakers, saying AI's takeover of childhood has happened in about three years, roughly twice as fast as social media.

What parents and teachers need to know: Don't wait for the "big AI talk." Start with one small question this week: "Have you used AI for anything lately?" Then listen, without flinching, even if the answer is something like a body question or a feelings question. The point of that survey isn't to scare you. It's to remind you the conversation is already happening, with or without you, and your kid would rather it be with you.

This week the Wall Street Journal reported on the spread of AI tools that turn ordinary photos into fake explicit images, which are increasingly being used to bully and harass kids and teens. The report notes that laws cracking down on these tools have lagged behind cases, and that Google disabled the "nudify" search term in its app store after a WSJ inquiry, with Apple also disabling such searches. Schools are still working out how to respond.

What parents and teachers need to know: We all need a reminder that silence does not protect kids. Say it, before it ever comes up: "If someone makes or shares a fake image of another person, that is harm and it isn’t your fault. If it ever happens to you or someone you know, come to an adult right away. You will not be in trouble for asking for help." Take it down can help pull these images down off social media and should be a resource for all kids and their parents.

This week's game builds the single most important skill from this week's stories: telling the difference between something that's friendly and something that's trustworthy. This game teaches discernment: friendly ≠ true ≠ safe (ties to the Common Sense Media, UNICEF, and chatbot stories).

What You Need: no supplies needed

How to Play

Sample statements to read:

Closing question: "What's the difference between something being nice to you and something being good for you?" Let them sit with it and share their answer.

Why it works: Kids already know the difference between someone who's nice to them and someone who's good for them, they feel it on the playground every day. This game moves that instinct over to AI, where "friendly," "true," and "safe" get blurred together. Sort those three apart out loud a few times, and kids start doing it on their own.

I know some adults want to say, “No AI for kids.” I get it. I really do. But this is not a plan. Pretending AI is not here will not protect kids. It just means they may end up learning about it from the internet, from friends, or from the tool itself. The laws matter. School policies matter. Better parental controls matter. I want all of that. But those things are moving slower than the kids are.

Share this with another parent or teacher if it helps. Bring one of these stories to your school, your PTA, or your group chat. But at minimum, start the conversation.

In a world of no clear rules, thank you for taking AI literacy seriously.

See you all next Tuesday.

Amber Ivey (AI)