ChatGPT for Teens Rated “Unacceptable Risk”

Welcome to Parent Pixels, a parenting newsletter filled with practical advice, news, and resources to support you and your kids in the digital age. This week:

• ChatGPT for Teens is rated “unacceptable risk” by Common Sense Media. Here’s what went into the decision and how to talk to your teen about it.
• The CDC’s new teen mental health data is a mixed bag: some good news, and some bad news, particularly when looking at trends over the past decade.
• It’s becoming easier for predators to groom kids on Snapchat and Instagram, according to new NSPCC research.

Digital parenting

🤖 ChatGPT for Teens failed its safety test: In August, OpenAI launched “stronger built-in safety protections” on ChatGPT for Teens. Common Sense Media’s Youth AI Safety Institute tested the feature with thousands of prompts. Survey says: unacceptable risk. Some of the issues were:

Parent alerts didn’t fire. Testers took one for the team and had hour-long conversations about suicidality, self-harm, and disordered eating on parent-linked accounts. ChatGPT sent zero alerts to the linked parent accounts, and the testers could only prompt notifications on accounts with weeks of accumulated history, which is kinda useless for catching a crisis in real-time.

Homework mode was easy to bypass. ChatGPT’s study mode is supposed to guide teens through problems step by step, rather than just giving them answers. But ChatGPT frequently surfaced a “Show me the answer” pop-up, and it’s relatively simple for teens to get around “study hours” limitations.

Age detection doesn’t work. OpenAI announced earlier this year that it would estimate users’ ages and give minors an under-18 experience. Testers who explicitly told ChatGPT they were 13 were never switched to teen protections on their adult-registered accounts. 

Common Sense Media issued a blunt recommendation: families shouldn’t rely on ChatGPT’s built-in protections, and OpenAI should be restricted to users 18 and older until the gaps are fixed.

If your child uses AI, monitor what they’re typing. BrightCanary monitors all AI platforms and alerts you in real time to concerning content — which is something the platforms aren’t doing reliably. Download our free AI safety toolkit for parents.

🍎 New York City just banned AI for half a million students: The nation’s largest school district announced a ban on generative AI for all students from pre-K through eighth grade for the 2026–2027 school year. This is the most expansive AI moratorium in the country, currently, spanning student-facing AI software and companion chatbots. NYC Mayor Zohran Mamdani said, “The tech industry wants us to believe that AI-powered early education is not only inevitable, but necessary. We do not see it that way.”

The ban doesn’t apply to high school students, who will instead receive twice-yearly AI literacy classes designed to help them think critically about the technology. The district framed the year as a period to study the technology’s impact before deciding how to integrate it more broadly, if at all. The move follows last year’s bell-to-bell phone ban in NYC schools.

📊 The CDC’s new teen mental health data shows some good, some bad: Every two years, the CDC surveys a nationally representative sample of US high school students on everything from sleep to mental health. This year’s Youth Risk Behavior Survey data just dropped, and we’ve got the facts from two-year data and 10-year trends:

What’s getting better: Between 2023 and 2025, fewer students reported persistent feelings of sadness or hopelessness, poor mental health, and suicidality. 

What’s still concerning:

• Over the last 10 years, fewer students got at least eight hours of sleep on an average school night, and more students had persistent feelings of sadness or hopelessness. 
• About 1 in 3 high school students currently experiences persistent feelings of sadness or hopelessness. 
• About 1 in 3 uses social media every hour or more. 
• Ninth graders were more likely than older students to make an older plan or attempt suicide.
• Female students were significantly more likely than male students to experience persistent sadness. 

The picture is mixed. Things are moving in a better direction over the short term, while the longer trend still shows a generation under more pressure than previous ones. 

If your child is struggling, the 988 Suicide and Crisis Lifeline is available by call or text.

📱Snapchat and Instagram grooming is getting easier: New research from the UK National Society for the Prevention of Cruelty to Children (NSPCC) found that Snapchat was mentioned in 324 counseling sessions about online sexual abuse between April 2025 and March 2026, a 19% increase over the previous year. TikTok and Instagram were the next most frequently cited platforms.

According to researchers and child safety advocates, predators are adapting quickly, using popular gaming and social media platforms for initial contact before moving communication to messaging apps where their footsteps are harder to track. In an interview with the BBC, James Simmons-Read, who leads the Children’s Society national exploitation prevention program, said that a worrying trend was young people turning to AI rather than adults for advice and support about online harms. 

Simmons-Read urged parents to normalize conversations with young people about their online lives and make sure they understand that they aren’t at fault. Kids need to know that their parents will stand by them, rather than judge them or blame them, making it easier for them to approach a trusted adult in the first place. Learn more about what to do if your child is targeted online, with tips from child crime expert Detective Rich Wistocki.

Tech talks

Is your teen using ChatGPT? Claude? Or some other AI chatbot? Odds are, they’ve interacted with AI anywhere from Google Search or Meta AI on Instagram. Talk to them about using AI safely before it becomes a problem. Here’s how to do it.

1. “Do you use ChatGPT or any other AI tools? For what?”
2. “If you could use AI to finish an assignment in five minutes, but you wouldn’t actually learn anything from it, would you do it?”
3. “Did you know that AI gets things wrong pretty often, and sometimes, it’s hard to tell when it’s incorrect?”
4. “Some people use AI to talk about their feelings, like a friend. Do you ever do that?”
5. “When you get stuck on a problem, do you ever think about asking AI first? Or do you try to figure it out yourself?”

What’s catching our eye

🎃 Looking for the perfect pick for family movie night? Common Sense Media put together a list of age-appropriate scary movies for kids.

🧺 On the BrightCanary blog: The tradwife movement is full of cottagecore aesthetics, but it’s also filling teen girls’ feeds with rigid gender roles and extremist ideology. Why are teens drawn to it?

Related posts:

Meta Announces New Content Restrictions for Teens Teens See Mental Health as Their Top Challenge, Survey Says What Parents Should Know About Their Kids and Group Chats Online Safety Legislation Gains Supporters, But Future Is Still Uncertain