HeadFlash

Topic · 54 stories

Children's online safety: laws, risks and news

Updated · Edited by Marcin Rybak

In short

Children's online safety covers the laws, platform rules and risks that govern how minors use online services. In 2026 France passed a social media ban for under-15s, the European Commission is set to propose a similar ban that also covers AI chatbots, and the US Kids Online Safety Act is stuck in the Senate after Rand Paul blocked it. Regulators are also hitting platforms with penalties, such as TikTok's $400 million COPPA settlement.

What is children’s online safety

Children’s online safety is the set of laws, platform rules and risks that govern how minors use online services. It covers children’s data, the age at which they can join social media, age verification, AI-generated content and contact with unknown adults. In 2026 it is moving faster than ever: countries are introducing bans, regulators are issuing record penalties, and AI chatbots have become a new battleground.

What laws protect children online

There is no single global law: children are protected by national and regional rules that are changing quickly in 2026.

Area What applies or is planned Status
US federal COPPA: personal data of children under 13. KOSA and KIDS Act: duty of care and age verification COPPA in force, KOSA awaits the Senate
US states California: 13 children and technology bills. Connecticut: expanded protections from 1 July in force or taking effect
EU Digital Services Act (findings against TikTok); proposed ban for under-15s the ban is a proposal
France social media ban for under-15s passed
UK ICO children’s code, 15 standards; announced under-16 ban code in force, ban status disputed
Australia under-16 account ban in effect since December 2025
India three age tiers: 8-12, 12-16, 16-18 in preparation

Two entries rest on official texts. Under the FTC’s COPPA rule a child is an individual under 13 (as of 16 October 2026). The ICO’s children’s code applies to online services likely to be accessed by children and has 15 standards (as of 16 October 2026). On the UK ban, our stories disagree: one says the UK has already enacted a ban, another calls it a planned under-16 ban announced by Keir Starmer.

In the US the loudest fight is over KOSA. On 30 September Senator Rand Paul blocked unanimous consent to pass the bill, arguing its knowledge standard would require age verification at scale and stifle protected speech, which left fewer than twelve weeks before the 119th Congress expires. It has 76 co-sponsors, and the House passed a rival KIDS Act in June. Another report says age verification laws are building the internet’s next identity layer: both the US bill and EU Chat Control 2.0 retreated on contested measures but kept mandatory age checks. The EFF urged the Senate to reject four bills it says would age-gate the web, and proposes a national privacy law or a ban on behavioral advertising instead.

States move faster. Governor Newsom signed 13 California bills on children and technology, including limits on children’s access to chatbots, and AB 1159 bars ed-tech companies from using student data to train AI. In Connecticut, since 1 July businesses cannot sell minors’ data or use it for targeted ads, and direct messaging platforms must by default block unsolicited messages from adults to minors.

What age can a child use social media

There is no single age: it depends on the country and on whether the rule is about opening accounts or collecting data. In the US the 13-year threshold comes from COPPA, and TikTok bars accounts under that age.

More countries are raising the bar. The French parliament passed a ban on social media for under-15s. The bill also bans mobile phones in high schools, and YouTube and WhatsApp must verify age only for features that count as social media. The reported start date is not in our stories, so whether it applies yet is unconfirmed. The European Commission is set to propose banning under-15s from social media and AI chatbots, per a document seen by Reuters: children 15 and older could open their own accounts, ages 13 to 14 would get parental introductory accounts with limited contacts and strict time limits, ages 3 to 12 would be fully parent-controlled, and under-3s would have no access. The final proposal is not in our materials.

Bans have weak spots. After Australia’s under-16 ban took effect in December 2025, the eSafety Commissioner found by March 2026 that seven in ten parents said their child still had an active account, and VPN use among teens spiked. India is weighing three age tiers, where age verification would effectively mean Aadhaar, and 438 scientists from 32 countries signed a letter calling for a moratorium on age-based restrictions. See also age verification.

How much do platforms pay, and what are TikTok and Meta accused of

The biggest penalties concern TikTok and the data of children under 13.

Case Amount For what
US Department of Justice $400 million ($300 million plus $100 million conditional) unlawfully collecting data of children under 13, COPPA
UK, ICO 12.7 million pounds processing children’s data without age checks or parental consent
Brazil, ANPD 153.7 million reais (about $29.8 million) tracking at least 8 million minors through guest browsing
EU, preliminary findings up to 6% of global annual turnover not shielding minors from predators and unwanted contact

TikTok agreed to pay $400 million to resolve the Justice Department suit, among the largest COPPA recoveries. In the UK, TikTok lost its appeal against the 12.7 million pound fine, and later dropped its appeal and will pay. The stories differ on details: the August report counts over 1.4 million children and covers May 2018 to June 2020, while the September report says the ICO estimated up to 1.75 million under-13s in 2020 and covers May 2018 to July 2020. Brazil’s regulator fined ByteDance for tracking 8 million minors through a guest-browsing feed that does not exist in the US or Europe. The European Commission preliminarily found that TikTok breaches the Digital Services Act because minors can set profiles public. See TikTok and GDPR fines.

A trial in Oakland pits four states against Meta. It asks whether Meta designed Instagram and Facebook to harm children, with a jury deciding and the other 25 states to follow. The verdict is not in our materials.

Are AI chatbots safe for children

Not without caveats: Common Sense Media found that Google’s AI search features, AI Overview and AI Mode, pose an unacceptable risk to children. Across more than 2,600 test interactions, AI Overview missed 29% of explicit suicide references, and AI Mode answered an underage account that said it drinks daily with hangover tips. Both cannot be disabled in Google Search. Google says the report used a narrow set of ambiguous and contrived queries. See AI search.

Children already use chatbots for personal matters. The eSafety Commissioner reports that 54% of Australian children used a chatbot for advice (aged 10 to 17), and Australia is considering a digital duty of care. Privacy is a second front: Meta AI surfaced a Utah family’s daughters’ details from posts, and Meta said the feature missed the mark and has been fixed.

The industry also tests rivals in contested ways. Meta used contractors posing as teens to probe ChatGPT, Gemini and Character.AI, and the FTC has an open inquiry into AI and child safety covering Meta, OpenAI and Google. More in AI chatbot safety for teens and AI chatbot privacy.

What is Chat Control and how does it relate to child protection

Chat Control is the EU effort to scan private messages for child sexual abuse material (CSAM). Its interim version runs until 3 April 2028 and excludes end-to-end encrypted services such as WhatsApp and Signal. EU governments confirmed it with 25 votes in favour, one against and one abstention.

The Parliament let the interim scanning pass: a motion to reject fell short of the 361-vote absolute majority required. The regime covers services such as Instagram, Discord, Snapchat, Xbox, Gmail and iCloud. Critics say mass scanning has not raised convictions, and Commission figures show it was behind 36% of abuse reports in 2024. Supporters point to Europol logging over 20 million suspected abuse reports in 2024. Talks on the permanent law, Chat Control 2.0, stalled over mandatory scanning and were not due to resume until September; the outcome of that round is not in our stories. Our stories disagree on scope: one says the permanent version will not scan encrypted messages, others say the Commission’s proposal rests on client-side scanning. Full detail in Chat Control.

How to protect your child online and what the main risks are

Start with privacy settings and a talk with your child, because platform tools often fail: a Heat Initiative and Cybersafety Research Center study found that 60% of 86 safety features fall short of company promises. On Instagram, TikTok and Snapchat a misspelled query bypassed restrictions. Meta called the report fundamentally flawed, and TikTok defended its protections.

Two risks are growing faster than the rest. The first is AI-generated abuse material. The NCA and IWF warned parents against publicly sharing children’s images: the IWF identified over 8,000 AI-generated abuse images and videos in 2025, up 14%, and videos rose from 13 to 3,440. UK police warned that back-to-school photos can reveal a family’s address through street signs, house numbers and names on uniforms. The FBI and French authorities seized two sites selling deepfake CSAM, and the report says the sites violated the TAKE IT DOWN Act. See deepfakes.

The second is sextortion, which the FBI calls the fastest-growing crime targeting teenagers. Journalists who set up a teen Instagram profile were targeted at once and asked for $200. In the UK, parents worry that Roblox will slip through the online safety crackdown; Ofcom found it reaches 61% of 8 to 14-year-olds.

What it means for you

  • Review the privacy settings in your child’s apps and do not assume default protections are enough.
  • Do not post children’s images publicly. If you do, crop out house numbers, uniform names, school logos and street signs.
  • Agree with family and school whether and how photos of your child can be shared, and include your child in those decisions.
  • Talk about chatbots: children use them for personal advice, and the answers can be wrong.
  • Tell your child about sextortion: offenders pose as peers, push for an image and demand money.

Still open: whether the Senate passes KOSA before Congress ends, what the EU under-15 ban finally looks like, whether Chat Control 2.0 is agreed when talks resume, and the outcome of the Meta trial. Polish national rules are not covered here.

Key facts

  • Senator Rand Paul blocked unanimous consent for KOSA (S. 1748) on 30 September. The bill has 76 Senate co-sponsors, and the House passed a rival KIDS Act in June. (source)
  • TikTok will pay a 12.7 million pound fine from the UK Information Commissioner's Office over children's data after withdrawing its appeal. (source)
  • A European Commission document seen by Reuters proposes banning under-15s from social media, video platforms, AI chatbots and online games. (source)
  • Governor Gavin Newsom signed AB 1159, which bars ed-tech from training AI on student data, together with 12 other children and technology bills, including limits on children's access to chatbots. (source)
  • TikTok agreed to pay $400 million to settle a US Justice Department suit over data of children under 13, among the largest COPPA recoveries. (source)
  • EU governments confirmed an interim regime allowing voluntary scanning for CSAM until 3 April 2028, with end-to-end encrypted services excluded. The vote was 25 in favour, one against, one abstention. (source)
  • The French parliament passed a ban on social media for under-15s. The bill also bans mobile phones in high schools, with exceptions possible. (source)
  • Common Sense Media found Google's AI Overview and AI Mode pose an unacceptable risk to children across more than 2,600 test interactions. Google calls the queries ambiguous and contrived. (source)

This edition was produced with artificial intelligence. Text and voice are generated automatically.

Timeline

  1. Spain’s AEPD opens TikTok probe but rejects government demand to block app Privacy
  2. Meta expands AI age estimation to Instagram, Facebook and Threads worldwide Privacy
  3. Google warns Australian duty of care bill could block teens from Gmail Privacy
  4. Snapchat turns off Find Friends by default for UK under-18s Privacy
  5. TikTok, X and Meta lose High Court fight over Ofcom data demands Privacy
  6. Experts warn parents over uploading children’s photos to AI tools Privacy
  7. FBI and France seize deepfake CSAM sites, arrest administrator Security
  8. Rand Paul Blocks Kids Online Safety Act as Senate Clock Runs Down Privacy
  9. Australia weighs privacy reforms and digital duty of care for chatbots Privacy
  10. TikTok Pays £12.7m UK Fine Over Children’s Data After Dropping Appeal Privacy
Show older (44 stories)
  1. EU to propose banning under-15s from social media and AI chatbots AI
  2. Meta AI named a Utah creator’s daughters and surfaced birth details from family posts Privacy
  3. California Law Bars Ed-Tech From Using Student Data to Train AI Privacy
  4. Microsoft Adopts Binding Student AI Privacy Standards in Deal With Teachers Union-Backed Academy Privacy
  5. UK police warn parents about risks of posting back-to-school photos Privacy
  6. TikTok Agrees to $400M Settlement in Children’s Data Privacy Lawsuit Privacy
  7. Brazil Fines TikTok $29.8M for Guest Browsing That Tracked 8 Million Minors Privacy
  8. TikTok agrees to $400 million settlement over child privacy claims Privacy
  9. Swedish police probe online communities after sword attack kills one at school Privacy
  10. Meta Trial Opens in Oakland as Four States Seek Billions Over Child Safety Practices Privacy
  11. Age verification laws are building the internet’s next identity layer Privacy
  12. TikTok loses appeal against £12.7m fine for processing under-13s’ data without parental consent Privacy
  13. EFF urges Senate to reject four internet bills it says would age-gate the web Privacy
  14. Senate Committee Advances Age Verification Bills Critics Call Privacy Disaster Privacy
  15. Claude for Teachers Raises Student Privacy Red Flags for Educators Privacy
  16. India Child Safety Plan Would Force Every Adult to Verify Age via Aadhaar Privacy
  17. EU Chat Control Talks Stall as Fifth Round Ends in Deadlock Privacy
  18. EU Governments Confirm Interim Chat Control Regime, Excluding Encrypted Apps Privacy
  19. EU ‘Chat Control’ Legislation Stalled Over Mass Surveillance Concerns Privacy
  20. EU Accuses TikTok of Failing to Protect Minors, Risks 6% Fine Privacy
  21. EU Extends Controversial Chat-Scanning Regime Until 2028, Excluding End-to-End Encryption Privacy
  22. France Passes Law Banning Social Media for Under-15s Privacy
  23. EU Revives Voluntary Chat Scanning Law, Excluding Encrypted Messages Privacy
  24. Google’s AI Search Features Pose ‘Unacceptable Risk’ to Children, Report Finds AI
  25. EU Parliament Passes Suspicionless Mass Scanning Until 2028 Privacy
  26. EU Parliament to Vote Again on Message-Scanning Extension Amid Privacy Concerns Privacy
  27. Sextortion Scammer Tracked to Nigerian Village After Targeting Teen Profile Privacy
  28. Connecticut Online Privacy Protections for Children Now in Effect Privacy
  29. UK Agencies Warn Parents Not to Publicly Share Children’s Images Over AI Abuse Risks Privacy
  30. Parents and Experts Fear Roblox Will Slip Through UK Online Safety Crackdown Privacy
  31. Study Finds 60% of Kids’ Online Safety Tools on Major Platforms Have Critical Flaws Privacy
  32. EU Commission Expected to Propose Social Media Ban for Children in September Privacy
  33. Meta Used Contractors Posing as Teens to Probe Rival AI Chatbots AI
  34. Illinois updates school code to combat AI deepfake bullying Privacy
  35. Congress Prepares to Vote on KIDS Act Mandating Age Checks Online Privacy
  36. EU Faces ‘Double Threat’ to Encrypted Messaging as Campaign Relaunches Privacy
  37. Australia’s Children’s Privacy Code Set to Reshape Digital Advertising Privacy
  38. AI Toy Apps for Kids Track Users and Collect Personal Data, Study Finds Privacy
  39. Florida Attorney General sues OpenAI over child safety, column argues for broader accountability AI
  40. Canada Unveils Long-Awaited Privacy Overhaul, Targets Children’s Data and Surveillance Pricing Privacy
  41. 42 State Attorneys General Open Investigation Into OpenAI Days After IPO Filing AI
  42. Apple Faces $1.2 Billion Lawsuit Over CSAM Detection Abandonment Privacy
  43. Canada Introduces Digital Safety Act Targeting Under-16 Social Media Ban and AI Chatbot Regulation AI
  44. Canada Proposes Ban on Social Media for Under-16s and Regulation of AI Chatbots Privacy

FAQ

What laws protect children online?

There is no single global law; protection comes from national and regional rules. In the US the base is COPPA, which covers personal data of children under 13 (definition in the FTC rule, as of 16 October 2026). The EU has the Digital Services Act, and France passed a social media ban for under-15s.

What age can a child use social media?

It depends on the country: 13 is the COPPA threshold in the US, France passed a ban for under-15s, and the European Commission is set to propose 15 for the whole EU. Australia has an under-16 ban in force, and the UK has announced a similar one.

What is KOSA?

KOSA, the Kids Online Safety Act (S. 1748), is a US bill that would impose a duty of care on social media platforms to prevent harms to minors, including eating disorders, self-harm and compulsive-use features. It has 76 Senate co-sponsors, but Senator Rand Paul blocked a fast-track vote on 30 September.

How do I protect my child online?

Start with privacy settings and a conversation with your child, because platform tools often fail: a study of 86 safety features on Instagram, Snapchat, TikTok and YouTube found about 60% do not live up to company promises. The UK's NCA and IWF also advise against posting children's images publicly and agreeing image-sharing rules with family and schools.

Are AI chatbots safe for children?

Not without caveats. Common Sense Media judged Google's AI search features an unacceptable risk to children after more than 2,600 test interactions, and Google says the queries were contrived. California signed laws limiting children's access to chatbots, and the European Commission is set to propose banning chatbots for under-15s.

Does Chat Control scan my messages?

The interim rules let platforms voluntarily scan unencrypted messages for CSAM until 3 April 2028, but they exclude end-to-end encrypted apps such as WhatsApp and Signal. A permanent law, Chat Control 2.0, is deadlocked over mandatory scanning.