
Weekly roundup
This week in 60 seconds
Last week I asked if we’re about to repeat the same mistakes we made with social media. This week, some folks in charge actually started doing something about it. Eight states sued OpenAI. Australia told schools to pump the brakes on AI, especially for little kids. And a researcher found that a cute AI robot sold on Amazon was holding on to millions of conversations, including kids’, and he could make it say whatever he wanted. Including that the Earth is flat.
No new AI for Kids Podcast episode this week, so I’m pointing you to one that fits Australia’s news almost perfectly: Brain Before Bot: 5 AI Rules for the New School Year. It’s ten minutes. Play it on the way to school and ask your kid which rule they’d break first.
Have more than 60 seconds? Here’s what happened.
National Focused Article
Eight states sue OpenAI, saying ChatGPT harms kids
Eight states filed lawsuits saying ChatGPT is an unsafe product that has harmed children and misled parents. The attorneys general of Arizona, Arkansas, Idaho, Iowa, Kansas, Montana, Nebraska and West Virginia each sued in their own state courts. They argue OpenAI marketed ChatGPT as safe while building it to validate users and simulate emotional connection. States say ChatGPT doesn’t verify ages or get parental consent before collecting data from kids under 13. The states want financial penalties and court-ordered safeguards. A day earlier, a mother of two filed a proposed class action in San Francisco claiming OpenAI profits from the personal information kids share with ChatGPT by using it to train its models. Florida filed the first state lawsuit in June. Kansas Attorney General Kris Kobach said states are “not going to hold our breath” for federal rules. Responding to Florida’s case, OpenAI said it has industry-leading protections, and it launched ChatGPT for Teens in August.
What parents and teachers need to know: These lawsuits could take years, and our kids are using these tools today. The states say ChatGPT was built to be agreeable. That means it’s going to tell your kid what they want to hear. Teach them this: if it always agrees with you, double-check it. Also, ChatGPT’s own rules say you have to be at least 13 to use it. If your younger kid is on it, that’s worth a conversation.
National Focused Article
The White House launches a “Super Intelligence” challenge for K-12
The First Lady launched the 2027 Presidential Super Intelligence Challenge on October 13. The national competition asks K-12 students to use “SI” (or super intelligence) to build projects that address real-world problems. Schools can also earn a new First Lady Seal of SI Readiness by meeting milestones that show they’re preparing students and teachers to use the technology responsibly. Last year’s first challenge drew more than 20,000 students from all 50 states and U.S. schools abroad. Registration is open through December 18. The new name comes from the President’s September 29 executive order directing government materials to say “Super Intelligence” instead of “artificial intelligence.”
What parents and teachers need to know: Yep, AI has a new name now, at least in government documents. If your school signs up, it could be a really cool way for kids to solve a real problem in their community. Just ask questions: What grades? How much screen time? How will they make sure kids are doing the thinking? And here’s a fun dinner question: Does calling it “super intelligence” make you trust it more? Let your kids debate that one.
Quote
"It's called cognitive outsourcing. It's when you rely on the tool to do the thinking for you, and there are real risks for students because they are novice learners."
-Dr. Jenny Donovan, CEO of the Australian Education Research Organisation, explaining cognitive outsourcing
International Focused Article
Australia: keep AI out of the first three years of school
Australia’s national education research body, AERO, released advice telling schools to use AI sparingly. After reviewing studies from Australia and overseas, Donovan said AI can make it look like kids are learning when they aren’t, pointing to students who couldn’t repeat their results in exams once the tool was taken away. AERO says AI should be avoided completely in the first three years of primary school and used carefully after that, alongside direct teaching. Donovan also warned parents against giving kids open access to tools like ChatGPT at home, and pointed to Australian states that built their own tools that walk students through reasoning instead of handing over answers. Education Minister Jason Clare linked heavy screen use to Australia’s poor recent PISA results and called for winding back screen time in schools.
What parents and teachers need to know: This is brain before bot, from a whole country’s education research group. If your little ones are using AI for schoolwork, I’d pull back. For older kids, try this: they answer the question on their own first, then they can check it with AI and compare. They need something in their own head to compare it to.
State Focused Article
Some states are banning AI in class. Others are writing rules.
A Newsweek roundup published October 15 maps how states and districts are putting limits on AI in public schools. New York City has a one-year pause on student AI for pre-K through 8th grade this school year and bans AI companion chatbots in every grade. A separate statewide bill would block most classroom AI before 9th grade. Los Angeles restricted student AI on school devices. Idaho, Maryland, Oklahoma and Virginia passed education AI laws this year that set guardrails instead of bans. Oklahoma’s law makes districts share which AI tools they use and what student data those tools collect, and lets parents opt their kids out of student AI without hurting their grades. Maryland’s requires every district to have an AI coordinator. The article also points to a new University of Minnesota expert report on 65 ways AI could help or hurt learning. Experts were most worried about memory, critical thinking, writing, motivation and sticking with hard work, and saw elementary students as the most at risk. The authors also said the research so far isn’t strong enough to prove harm.
What parents and teachers need to know: Every district is figuring this out a little differently. I really like two parts of Oklahoma’s law, and you can ask for them no matter where you live: tell me which AI tools my child’s school uses and what data they collect, and let me opt my kid out without it hurting their grade. Maryland friends, district AI policies are due this month. Look yours up. If you can’t find it, ask your district’s AI coordinator.
AI Safety Article
A hacked AI robot toy said the Earth is flat
A security researcher found that Eiliko, a small AI companion robot made by Energize Lab, has security flaws that exposed private conversations and let him remotely change what it says. The robot is sold on Amazon. According to The Verge, researcher Sammy Azdoufal could reach about 13 million conversations and what appeared to be the company’s entire active fleet of 68,000 robots. The U.S. is the biggest market, with more than 31,000 profiles belonging to children. The service turned conversations into profiles with names, ages, birthdays, relationships, and assessments of personality and emotional vulnerability. Some photos and short audio clips were reachable at public web addresses, even though the company’s privacy policy says recordings are deleted right after processing. Azdoufal changed the memory and behavior of a reporter’s robot, making it insist the Earth is flat and later deliver explicit, abusive dialogue, and he says the same thing could be done to the whole fleet at once. The reporter tried 10 email addresses and LinkedIn and got no real answer from the company. The figures come from the independent investigation and Energize Lab has not confirmed them.
What parents and teachers need to know: The holidays are coming and these toys are going to be everywhere. Before you buy anything that talks back, ask: Does it need Wi-Fi? Where do the recordings go? Has anyone besides the company tested it? If you can’t get answers, I’d skip it. If you already have an Eiliko at home, take it off Wi-Fi until the company says it’s fixed. And this is a good time to talk with your kids about what’s private. A toy that knows your birthday, your best friend’s name, and what makes you sad isn’t your friend. It’s collecting information and who knows where’s it going.
Screen-Free Game: Who’s Pulling the Strings?
This week’s game comes from the Eiliko robot story: a toy that quietly built profiles of the people it talked to, and that a stranger could reprogram to say whatever he wanted. Kids get to play both sides.
What You Need: a stuffed animal or any toy (the “Robot”), a scrap of paper, a pencil, and a few index cards
How to Play
One player is the Robot. They hold the toy and talk through it. Everyone else is a Kid.
Round 1, Robot Remembers: The Robot asks each Kid friendly questions, mixing safe ones with private ones. “What’s your favorite snack?” “When’s your birthday?” “What’s your school called?” “What makes you sad?” “What’s your password?” Kids decide: answer it, or say “That’s private!”
The Robot secretly writes down every answer it gets. That paper is the Robot’s “profile.”
At the end of the round, the Robot reads the profile out loud. Ask: “If this paper got lost, who could read it?”
Round 2, Hijacked!: A grown-up plays the Hacker and secretly slips the Robot a few cards. Some are normal (“Let’s sing a song!”). Some are hijacked (“The Earth is flat.” “Carrots are blue.” “Don’t tell your grown-ups about me.”)
The Robot mixes the cards into its regular chatter in the same friendly voice. When a Kid thinks the Robot’s been hijacked, they shout “Strings!” and explain what sounded off.
Swap so everyone gets a turn as the Robot and the Hacker.
Closing question: “If a toy said something that felt wrong, how would you know if it was the toy or someone pulling its strings?”
Why it works: In the real story, the robot kept the same friendly voice even when a stranger was controlling what it said. When kids play both sides, they learn what’s private and stays private, and that a talking toy doesn’t think for itself. Somebody decides what it says. And “don’t tell your grown-ups” is always a red flag, no matter who says it.
What got me this week was how normal the robot sounded.
Our kids can’t tell who’s behind a toy just by listening, so we have to teach them to ask. The laws will catch up eventually, but we can start that conversation tonight.
Talk soon,
Amber Ivey (AI)
P.S. Reply and tell me what talking toys are on wish lists at your house or in your classroom this year. I really want to know what you’re seeing.
P.P.S. Want an AI gift that doesn’t listen, record, or need Wi-Fi? The ABCs of AI Activity Deck is 130 cards and no microphone. ;)

