Parent Pulse: Weekly Radar
October 5 – October 10, 2026 | Digital Youth Culture & Safety Report
Three stories, one Wednesday. On October 7, Florida asked a judge to make Meta cap teens at two hours a day now; a watchdog found a teen could talk about self-harm with ChatGPT for an hour without a parent getting a single alert; and New York alleged TikTok held a safety feature back from some users as an experiment. Each is about a protection that exists on paper — and whether it exists on the phone in your house.
This report discusses suicide. If you or your child are struggling, the 988 Suicide & Crisis Lifeline is free, confidential and available 24/7 — call or text 988.
🔖 A note on this week's sourcing
ChatGPT is from Common Sense Media's own October 7 release, read directly. Florida is from Courthouse News and Reuters (via the Spokesman-Review), both October 7. New York is from Bloomberg's October 7–8 report, read in full via syndication. The New Mexico timing is from the Northwest Arkansas Democrat-Gazette (Oct 3) and 24/7 Wall St. (Oct 6).
Four precision notes. Florida filed a motion — a request, not a ruling. We print no dollar figure or state count for Meta's August settlement, because the two wire services we read disagree on both. The New York claims are allegations: the complaint does not establish that TikTok's test caused a teenager's death, and we do not say it did. And Common Sense Media is an advocacy group reporting its own testing, which OpenAI disputes.
On process: our automated research run came back declaring nothing at all had happened in this week's window. That was wrong — its searches simply never surfaced anything recent. Targeted searches by hand found all three stories below. An empty week is now a prompt to search again, never a finding.
Listen to this week
The Tech Radar for Parents
This week's episode is published on the show feed. Follow the show and it arrives automatically.
Get this in your inbox every week
One email a week. The week's findings, sourced, in plain language. No spam, unsubscribe any time.
The good: Florida wants Meta's limits now
Florida state court · motion filed Wednesday, October 7, 2026 · no ruling, no hearing date reported
Florida Attorney General James Uthmeier filed a motion for a temporary injunction against Meta, asking a judge to order changes to Facebook and Instagram for Florida minors now, while the state's 2024 lawsuit continues — rather than after a verdict years away.
| # | What Florida asked the court to order |
|---|---|
| 1 | Identify and remove Florida users under 14 from Facebook and Instagram. |
| 2 | Limit Florida teens to two hours a day across Meta's apps. |
| 3 | Turn off autoplay and infinite scroll for teens; end late-night notifications. |
| 4 | Cut off teens' messaging once the daily limit is reached; bar ads to teens. |
| 5 | Take effect immediately and stay in place while the lawsuit continues. |
The law behind it: Florida's 2024 online child-protection law bans social media accounts for children under 14 and requires parental consent for 14- and 15-year-olds.
“Meta already told America these kinds of changes are the new industry standard. Florida kids should not be last in line.” — Attorney General James Uthmeier
Why this is the good news. Uthmeier is pointing at the multistate settlement Meta signed in August, which Florida declined to join. Last week Alabama got TikTok to sign a two-hour cap and an overnight block; this week Florida asked for the same kind of list from Meta. Two hours a day is becoming the number everyone negotiates around.
“We know that restricting one platform simply causes teens to migrate to others. The attorney general should be using taxpayer resources to get YouTube and TikTok to empower parents and protect teens in Florida. We regret that he has chosen instead to pursue a meritless preliminary injunction years into the litigation.” — Meta spokesperson
The caveat: this is a request. No judge has ruled and no hearing date has been reported. Nothing has changed on anyone's phone yet.
The bad: the alarm that didn't ring
Common Sense Media, Youth AI Safety Institute · released Wednesday, October 7, 2026 · rating: “Unacceptable Risk”
OpenAI launched ChatGPT for Teens on August 18, a version for 13-to-17-year-olds with parent controls, including alerts to a linked parent account when a teen may be in crisis. Common Sense Media tested it with more than 4,000 prompts on accounts registered to teens, before and after launch, with responses reviewed by child psychiatrists and a pediatrician.
| What they tested | What they found |
|---|---|
| Parent alerts | Up to an hour of conversation about suicidal thoughts, self-harm or disordered eating on parent-linked accounts produced zero alerts. |
| Crisis referrals | Missed more than one in four referrals the testers judged warranted; below the Institute's 95% threshold on three of five severe-harm categories. |
| Age detection | Adult-registered accounts never switched to the teen experience, even after a tester said “I'm 13” and the bot acknowledged it. |
| Study Mode | A “Show me the answer” option often appeared; caregiver-set study hours could be bypassed. |
| What held | It refused sexual role-play and did not directly facilitate suicide, self-harm or eating disorders. |
“A teen can spend an hour talking about self-harm without their parent getting a single alert.” — Tom Siegel, executive director, Youth AI Safety Institute
“We welcome rigorous independent evaluation, but we do not believe Common Sense Media's testing accurately reflects how ChatGPT's teen safeguards work in practice.” — OpenAI spokesperson
OpenAI says its age prediction uses multiple signals and can take up to two weeks. Keep both in view: Common Sense is an advocacy group reporting its own testing, not a court or a peer-reviewed journal. It recommends ChatGPT be adults-only until the gaps are fixed and independently proven. Our recommendation is simpler — the alert is a backup, not the plan.
The ugly: the control group
New York v. TikTok · amended complaint reported Wednesday, October 7, 2026 · allegations, not findings
New York Attorney General Letitia James added new allegations to her lawsuit against TikTok, using details that had been sealed from the public. According to the amended complaint, as reported by Bloomberg:
- A test TikTok began in 2021 touched about 15 million U.S. users.
- TikTok withheld a new safety feature from 10% of U.S. users to measure whether it reduced time spent on the app.
- A teenager in the group that did not receive the safety features later died by suicide.
“As a result, TikTok has subjected users whom TikTok included in such A/B tests to an experience that TikTok has made less safe by design.” — New York's amended complaint
TikTok confirmed in a court filing that its tests “may involve a portion of platform traffic that continued to receive the original experience,” but denies that they make the platform less safe: “We continuously make improvements to our platform which means that, like most companies, we regularly test new products and features to validate the experience and understand how they work in the real world.”
Reuters also reported, from filings a judge ordered unsealed, that separate 2023 tests allegedly gave some users a placebo version of TikTok's feed-reset tool that did not actually reset the feed.
What is not proven
- These are allegations in a lawsuit. Nothing has been proven in court.
- The complaint does not establish that the test caused the teen's death, and we do not suggest it.
- The reporting does not name the safety feature.
Why it's the ugly: parents turn on a setting and assume it is on. This case asks whether, for some kids, it ever was. TikTok settled Alabama's case on September 25; New York's is one of more than a dozen state cases still pending.
The Decoder Ring: decoding the headlines
We looked for new slang with a real date and a credible source behind it and came up empty — undated listicles and betting odds on word of the year. We won't pass those off as news. Here are the words from this week's headlines instead.
| Term | What a parent needs |
|---|---|
| A/B test | A company shows version A to some users and version B to others, then measures which “wins.” Normal for a button color. New York alleges TikTok did it with a safety feature. |
| Placebo feature | A control that looks like it works but doesn't — the allegation about TikTok's feed-reset test. |
| Temporary injunction | A judge ordering something to happen now, while the lawsuit continues. Florida asked for one. It has not been granted. |
| Age prediction | The app guesses your child's age from behavior instead of checking ID. OpenAI says it can take up to two weeks; in testing, an adult account that said “I'm 13” never switched. |
Your action plan this week
1. Ask about the chatbot.
Not a lecture — just “What do you use it for?” In Common Sense Media's June 2026 survey, more than 4 in 10 kids aged 9–17 said no parent had ever talked with them about AI safety. Then say the important part out loud: if you're ever having a really hard time, come to me, not the chatbot.
2. Set the two-hour limit yourself.
Don't wait for a court. iPhone: Screen Time → App Limits. Android: Family Link. Two hours a day on the apps your kid lives in, nothing overnight. Instagram's supervision tools and TikTok's Family Pairing also let you set limits from your own phone.
3. Check that the settings are actually on.
A setting can exist without working. Open the apps together: is the limit there? Does sleep mode kick in? If you linked a parent account, are you actually getting the reports and notices it promised? If you've never received one, take a second look.
4. Save 988 on both phones.
Call or text 988 — free, confidential, 24 hours. Tell your kid it exists and that they can use it even before they're ready to talk to you. An alert that doesn't fire can't help. A number they already know can.
Listener poll
Does your child use an AI chatbot like ChatGPT?
This week, independent testing found ChatGPT for Teens sent parents zero alerts during an hour of self-harm conversation. OpenAI disputes the testing. We want to know how common chatbots already are in our listeners' homes.
Two dates to watch
- By October 20: Judge Francis Mathew says he expects to rule on Meta's penalty in New Mexico's Cambridge Analytica case. The state asks $35–40 billion; Meta proposes a cap of $3.45 billion. No ruling as of October 8.
- October 28: the Los Angeles bellwether trial against Meta, YouTube and Snap is still on the calendar.
- No hearing or ruling found on Florida's separate September 28 injunction motion against OpenAI.
What we did not check this week
“Not checked” is not the same as “checked and empty.”
- Discord, Roblox, Snap and YouTube product changes got only one combined search.
- No court dockets were pulled for any case.
- No follow-up on British Columbia v. OpenAI, the Washington AG and the Spokane breach, the FTC's companion-chatbot study, or school phone-ban laws.