Parent Pulse: Weekly Radar
September 28 – October 3, 2026 | Digital Youth Culture & Safety Report
Most weeks this report covers what somebody is thinking about doing. Not this week. Alabama's Attorney General walked out of a Montgomery courtroom with a signed list of things TikTok has to change about the app your child is holding. Plus a jury that found Facebook misled people on purpose 43,899,720 times, a district that got its grade book back without saying whose data was taken, and the most-searched slang word of the year — which means nothing at all.
This report discusses AI-generated sexual images of children. 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
Alabama is from the Attorney General's own PDF release of September 25, extracted and read directly. New Mexico is from Source New Mexico, NM Political Report, Courthouse News and the Albuquerque Journal — plus, for the first time, the judgment document itself from nmcourts.gov. Spokane is from the Spokesman-Review and KXLY and the Washington AG's breach directory. The research is JAMA Pediatrics (Sept 28) and The Brussels Times/Belga on Ghent University (Sept 25).
Five precision notes. Alabama receives a minimum of $100 million; the $300 million figure is a contingent ceiling, never the state's money. Alabama settled three days before trial — the release's own wording is counterfactual, so no state has taken TikTok to a jury. Spokane's staff-only scope is preliminary and unchanged. The NCMEC figure below came from a television interview, not a published report, and we say so. And the JAMA paper is an umbrella review — not a study, not a meta-analysis.
On process: for a fourth straight week our research pipeline's summary buried its own best material, this time declaring the Meta verdict — which it had itself called the likeliest major development — as having received no research pass. Four of its own agents had it in detail. We found it by reading the raw evidence rather than the summary, which is now the only way we use this tool.
Listen to this week
Episode published 2026-10-05 · 30 min
Alabama Just Made TikTok Change the App. Here's the List
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: Alabama moved first, and published the list
Montgomery, Alabama · announced Friday, September 25, 2026 · five-year term
Alabama Attorney General Steve Marshall announced a settlement with TikTok Inc. The money, in the state's own words: “Alabama will receive a minimum of $100 million, due to the state within 45 days, with the potential to receive up to $300 million if certain conditions are met.”
Read that carefully, because it is already being reported wrong. Alabama receives a minimum of $100 million. The $300 million is a ceiling, not a cheque — it arrives only if certain conditions are met.
But the money is not why this matters. The settlement requires TikTok to change the product, and the state published the requirements. Verbatim:
| # | What TikTok must do |
|---|---|
| 1 | “A two-hour daily time limit, with parental controls that allow parents to further limit the amount of time their children spend on TikTok.” |
| 2 | “‘Productive pauses’ that interrupt teen users after 15 minutes of continuous use and again at 60 and 90 minutes to limit endless scrolling.” |
| 3 | “‘Nighttime access’ that restricts children's access from 12:00 a.m. to 6:00 a.m., plus restrictions on messaging and push notifications overnight and during school hours.” |
| 4 | “Robust age assurance measures to more effectively verify the age of young users.” |
| 5 | “Stronger content moderation provisions, including limits on discoverability of teen accounts by adults and parental notification of suspicious interactions between teen users and adults.” |
| 6 | “A blanket prohibition on the use of cosmetic filters by teen users.” |
| 7 | “A default non-personalized feed for teen users.” |
| 8 | “Stronger, more user-friendly parental controls.” |
Item 3 is the sleeper — restrictions on messaging and push notifications during school hours, a settlement quietly doing what dozens of states have spent two years trying to legislate. Item 6 is the one parents of teenage girls will notice: not a toggle, a blanket prohibition. Item 7 turns the algorithm off by default.
“This is a great day for Alabama parents. Tonight, they can rest easier knowing real protections are in place to shield their children from the dangers of social media addiction. TikTok has agreed to give parents real control over what their kids see and how much time they spend on the app.” — Attorney General Steve Marshall
The clever part: a bounty on copying
The gap between $100M and $300M is contingent on other state attorneys general reaching comparable agreements, paying out in tranches as ten, then twenty, then thirty, then forty other states sign on. Alabama wrote a template and attached a bounty to it. This is less a story about one state's cheque than about one state trying to start a cascade.
Four honest caveats
- These are obligations as described by one side's press office. The release states no effective date, no per-feature deadline, and TikTok is quoted nowhere in it.
- A trade summary of the decree suggests the two-hour cap may apply to the short-video feed, not the whole app, with Search, messaging, settings and longer videos exempt. We could not confirm that against the decree text.
- “Within 45 days” of what is not stated — from announcement, signing, or court entry. So we will not give you a date.
- Alabama did not beat TikTok in court. It settled three days before trial. The release says Alabama “would have been the first state to bring its claims against TikTok to trial.” Would have been. No state has. That is not a criticism — $100M and a product redesign is a good outcome — but a negotiated outcome and a verdict are different things.
The verdict: a jury said Facebook misled people on purpose, 43,899,720 times
Santa Fe, New Mexico · verdict Friday, September 25 · penalty hearing Thursday, October 1
After a two-week trial before Judge Francis Mathew, a jury found Facebook willfully violated New Mexico's Unfair Practices Act 43,899,720 times. Jurors weighed 29 statements the company made — about whether users could control their own data, about hate speech, about misinformation — and found 26 of the 29 false or deceptive.
This is not the child-safety case. New Mexico has two cases against Meta; this is the adult data-privacy one, rooted in the Cambridge Analytica harvest of roughly 87 million profiles.
The 44 million is an arithmetic construct, not a headcount. The jury assigned 2.1 million violations to each statement carried in major news outlets (roughly New Mexico's population) and 1.4 million to each made through company channels. Do not read it as 44 million people harmed.
| Side | Position at the October 1 penalty hearing |
|---|---|
| New Mexico (Randi McGinn) | $35–40 billion |
| Meta (Matt Nicholson) | $345 million – $3.45 billion, and likely to appeal |
| Statutory ceiling | ~$219.5 billion ($5,000 × 43.9M willful violations) |
The two sides are roughly a hundredfold apart. Meta argued Eighth and Fourteenth Amendment violations, said the state was “running up the score,” and argued New Mexico proved no harm to New Mexicans. Judge Mathew said he will rule by late October. Attorney General Raúl Torrez says civil-penalty money flows automatically to a fund supporting New Mexico schools and students.
And in the other case, we finally read the judgment
We have said for three weeks that only a docket pull would close this thread. This week we got the judgment itself — NM ex rel. Torrez v. Meta Platforms, No. D-101-CV-2023-02838, signed by Chief District Judge Bryan Biedscheid, e-filed August 6, 2026. The $942M is confirmed as $375M in jury-assessed penalties (75,000 violations × the $5,000 maximum) plus a $567M abatement fund.
But the money is not the interesting part. For five years, in New Mexico, the order also requires Meta to:
- maintain Teen Account features for under-18 users;
- build an under-13 prediction model;
- proactively request proof of age, and delete unverified suspected under-13 accounts within 30 days;
- treat a user as under 13 or under 18 when age cannot be precisely estimated;
- review the friend networks of deleted under-13 accounts;
- create a reporting portal letting school administrators flag suspected underage accounts;
- provide an age-assurance appeal process;
- report compliance to the court every June 30 and December 31.
A court order telling Meta to build a button for your principal. The court also held that COPPA limits its power to order age verification — the reason these orders stop where they do. And the five-year abatement clock is tolled during appeal only if Meta posts a supersedeas bond, which is why the pending $1.8B bond motion matters more than it sounds. Still no ruling on it, and no notice of appeal, in anything we checked.
The bad: ten days to the grade book, and still no answer
Spokane Public Schools, WA · staff systems Monday, Sept 28 · parent portals Wednesday, Sept 30
Two weeks ago Spokane went dark — a network security threat on Sunday, September 20; PowerSchool and payroll pulled offline; schools open, attendance on paper; and four days in, no restoration timeline at all. The update is genuinely good news: attendance and grading returned for staff and payroll came back on Monday, September 28, and the parent and student portals returned Wednesday, September 30. Ten days, with schools open throughout. Springfield, Massachusetts lost an entire week of school.
But the sentence we flagged as most likely to change has neither changed nor been confirmed. Re-checked October 2: “preliminary findings indicate only staff were affected” — still preliminary, neither revised nor reaffirmed. No official has used the word ransomware. No group has been named. The district brought in an outside technical expert and says it expects to learn more “in the coming days” — which was the posture two weeks ago too.
The leak-site entry for a masked education victim attributed to a group calling itself NightSpire is unchanged: still masked, still reading “Data is not available now,” nothing published, no deadline set. Still an unconfirmed criminal-group claim.
The date to actually watch
There is no breach notification from Spokane — or any Washington district — in the state Attorney General's directory. But nothing is due yet. The deadline lands around October 20. That is the useful takeaway for every parent reading this: circle your state's notification deadline, not the day the grade book comes back. Restoring systems is an IT milestone. The notification deadline is the legal one — the point at which a district has to stop saying “we are still investigating.”
The ugly: a few thousand to 180,000
Stacy Garrett, a vice president at the National Center for Missing & Exploited Children, says CyberTipline reports involving AI went from “a few thousand” in 2023 to more than 180,000 AI-generated-CSAM-related reports in 2025.
⚠ We could not verify this independently. It comes from a television interview, not a published report, and we tried and failed to get the underlying data. We give it to you attributed to the person who said it, with that caveat attached. If it holds, it is the most important statistic in this report.
On Henderson, Kentucky, where a middle schooler made AI nude images of two classmates: a sharpening, not a correction — it is three charges, four counts. The juvenile was detained Sept 21 and a juvenile-court hearing was held Sept 22; the outcome has not been reported anywhere. The district declined an interview, and its only family-facing notice is an annual sexual-extortion letter required by state law — not a letter about this.
Kentucky did not need a new law. It amended its statute in 2024 so matter portraying a sexual performance by a minor includes a computer-generated image, and prosecutors need not prove the depicted child exists. And Louisiana's 2026 Ivy Daniels Act requires schools to teach this from sixth grade up.
The research: two studies, pointing the same way
Ghent University (Friday, Sept 25) followed roughly 3,500 pupils across 20 secondary schools, measured before and after a full phone ban. Total daily smartphone use did not fall — pupils shifted it after school. 62% admitted using phones covertly at school anyway. No concentration improvement, higher restlessness, and more reported cyberbullying — though that last may reflect more reporting.
This is not the study we covered last week. That was one school, N=960, no control group. Twenty schools with pre/post measurement is a far stronger design, pointing the same direction.
JAMA Pediatrics (Monday, Sept 28) published an umbrella review — a review of 23 meta-analyses, covering 2,817,252 children in 64 countries. Screen use was associated with slightly lower academic achievement and language skills, and somewhat more emotional, behavioural and peer problems. It is correlational, the inputs are largely cross-sectional, and the effects are small.
Together: taking the phones away, by itself, did not fix what we were told it would fix. That is not an argument against phone bans — it is an argument against expecting a ban to do the work alone.
The Decoder Ring: the #1 word of the year means nothing
Source warning first. This ranking is not research — it is a commercial dataset from a word-game website, with no sample size, no search volumes and no methodology beyond “Google Search trends data.” The data also runs January to August, so it is the year's ranking, not this week's. A conversation starter, not evidence.
| Term | What a parent needs |
|---|---|
| 6-7 | The most-searched slang term of the year — and it means nothing. From a rap song title and a viral clip of a child at a basketball game. No definition, no hidden drug reference, no secret. The correct response is to let it go. |
| Bop | Used to mean a catchy song. Now a misogynistic insult aimed at girls, about promiscuity. If you hear a boy use it about a girl at school, that is not a music review. |
| Chud | A general-purpose insult — stupid, ugly, repulsive. |
| Mog / mogging | Dominating someone by being better-looking. The verb form of looksmaxxing, from the corner of the internet that teaches boys their face determines their life. The one to actually pay attention to. |
| Crash out | To lose emotional control. The only term appearing on both this list and Merriam-Webster's September additions — a decent sign it is real. |
The ratio is the lesson. The number one word means nothing; two of the other nine are insults. Most of it is noise designed to exclude you; a small amount is contempt pointed at someone. Telling those apart beats memorising any list — including this one.
Your action plan this week
1. Open TikTok with your kid and see what actually landed.
Check three things: is there a two-hour daily limit, and is it on; is there a midnight–6 a.m. block; is the teen feed personalised or not. If you are not in Alabama, use the numbers anyway — they came out of a negotiation where a company's lawyers had to concede they were reasonable. Screen Time and Family Link will do all three tonight, in any state.
2. Do the second half of the AI-images conversation.
You likely did the first half: it is a crime, children do get charged, delete it, don't forward it, tell an adult. The second half is for the child it happens to, and it is three sentences: it is not your fault. You are not in trouble. Tell me, and we will handle it. Then look up NCMEC's Take It Down — free, works on AI-generated images — tonight, not in a panic later.
3. If your district has had an incident, send one email.
Three questions, in writing, to the superintendent's office, copying your board member: which categories of student information were in the affected systems; what is the district's notification deadline under state law; and will students — not just staff — be offered credit monitoring? In writing matters, because a district that answers in an email has said something it must stand behind.
4. Ask your school what replaced the phones.
Not do we have a policy — you know the answer. What went into the time where the phones used to be? A full ban, measured across twenty schools, did not reduce total daily use; it moved it after school. If the answer is nothing, that is your answer.
What we did not check this week
We check five categories every week, and we tell you when we come up short. This week the list is long. “Not checked” is not the same as “checked and empty.”
- The Cat in the Hat AI threat trend — a third consecutive week with no in-window searching. The only source we reached predated our window. We cannot tell you whether it is escalating, cooling or mutating. The honest statement is that we stopped tracking it, not that it stopped.
- Discord — our lead story a fortnight ago — got no follow-up at all. Whether the age-sorting worked, whether people were misclassified, whether anyone complained: unknown. We told you it was coming and then did not go back and look.
- Five of eight platforms: Snap, YouTube/Google, Roblox, Character.AI and Steam. And TikTok's own newsroom still will not load for our tools — everything above about TikTok comes from a government source, not the company.
- Springfield, Massachusetts, second week running: the local-outlet sweep, the Massachusetts AG breach database, DESE, and whether Interlock published the data.
- British Columbia v. OpenAI — OpenAI's response, court assignment, RCMP or coroner developments.
- No court docket was pulled, so the New Mexico bond and appeal question stays open for a third week for exactly the reason it was open the first.
- Also unchecked: NetChoice v. Skrmetti post-remand; whether industry groups sued over California's 13 kids-and-tech laws; the FTC's companion-chatbot 6(b) report; sextortion/CSAM enforcement actions; viral-challenge injuries.
Four stories is what we could stand behind. It is not what we went looking for.