zoltner

Monthly summary

April 2026 summary: AI and child safety

The EU’s legal basis for voluntary scanning for child sexual abuse material lapsed in April, and governments tightened limits on children’s social media. Greece announced an under-15 ban from January 2027, Turkey’s parliament passed an under-15 restriction, and Roblox settled with Nevada, West Virginia and Alabama. In Washington, the Senate Judiciary Committee voted 22 to 0 for a bill that would bar minors from AI companions.

Europe tightened limits and lost a legal basis

The EU’s temporary exemption allowing platforms to detect child sexual abuse material voluntarily expired on 3 April after lawmakers failed to agree an extension. Thorn warned that files reported to NCMEC fell 58 percent during a seven-month gap in 2021, and Google said it would keep detecting such material voluntarily. The Commission announced the rollout of its EU age verification app on 15 April, adopted a recommendation on 29 April urging member states to make it available by the end of 2026, and preliminarily found that Meta’s Instagram and Facebook breach the Digital Services Act by failing to keep under-13s off. The allegations are unproven.

Greece’s government announced a ban on social media for children under 15 from 1 January 2027, with reliable age verification and re-verification of existing accounts. Turkish lawmakers passed a bill requiring platforms to block under-15s from opening accounts, pending the president’s approval, and French President Emmanuel Macron pushed for a blanket under-15 ban.

The UK argued over a ban

The House of Commons voted 256 to 150 on 15 April to reject a Lords amendment for a default under-16 social media ban, and on 27 April voted 272 to 64 to insist on its position while a minister said the government must act after its consultation. Prime Minister Keir Starmer told executives from Meta, Snap, Google, TikTok and X at Downing Street that things cannot go on like this, and the government tabled an amendment that would let ministers wait up to three years. The consultation had drawn nearly 50,000 responses, about 6,000 from children.

The government also proposed personal liability, with possible jail time, for executives whose platforms ignore Ofcom enforcement decisions on non-consensual intimate images, and new regulations took effect on 7 April requiring services to report child sexual exploitation content to the National Crime Agency. Ofcom opened investigations into Telegram and two teen chat sites. The allegations are unproven.

US courts, states and Congress

Massachusetts’ top court ruled unanimously on 10 April that Meta must face the attorney general’s suit over Instagram’s design, finding Section 230 does not bar claims aimed at Meta’s own conduct. Meta told a New Mexico court that the state’s proposed remedies would be too onerous and threatened to shut down its networks there. The allegations are unproven.

A 19-year-old man in Hampshire was jailed for grooming a 14-year-old he first contacted on Roblox. Roblox agreed to pay Nevada $12.5 million, West Virginia $11 million and Alabama $12.2 million, with commitments that include age verification before chat and limits on adult contact with under-16s. The allegations are unproven.

Florida’s attorney general opened an investigation into OpenAI over risks to minors and on 27 April extended it to a criminal inquiry covering the USF murders, and families of seven Tumbler Ridge shooting victims sued OpenAI on 29 April for failing to alert authorities. Senator Chuck Grassley asked eight technology companies to explain how they would improve reports of child exploitation. The allegations are unproven.

States kept legislating. Mississippi’s governor signed an age verification and parental consent law, Kansas signed Caleb’s Law against online sextortion including with AI-altered images, and Wisconsin’s governor signed five child internet safety bills. Nebraska passed a conversational AI safety law 49 to 0, Michigan’s Senate passed a companion chatbot bill, and California’s Assembly privacy committee advanced AB 1709 on an under-16 ban. Two representatives introduced a bipartisan bill on AI companion chatbots.

Australia, Asia and Canada

Australia’s government backed a Digital Duty of Care after its Online Safety Act review. Australia’s eSafety Commissioner reported about 4.7 million under-16 accounts removed, deactivated or restricted, while a Molly Rose Foundation survey of 1,050 Australians aged 12 to 15 found more than 60% of teens with pre-ban accounts still reached at least one platform.

Indonesia sent YouTube a letter of reprimand, its first sanction under the new rules. The allegations are unproven.

China issued interim measures on AI services that simulate human personality, barring inducement of emotional dependency, and a draft banning digital humans from offering intimate virtual relationships to under-18s. In Canada, Liberal Party members voted to back age limits on social media and AI chatbots, and Manitoba’s premier proposed restricting both for children.

Platforms, evidence and abuse imagery

Roblox announced Roblox Kids for ages 5 to 8 and Roblox Select for 9 to 15, and Meta expanded Instagram’s 13+ content setting for teens worldwide and later let parents see the topics teens discuss with Meta AI. OpenAI published a Child Safety Blueprint on AI-enabled exploitation, and Common Sense Media asked major AI companies to give $100 million each for independent research. A Drexel study of 318 Reddit posts found teens described addiction-like reliance on Character.AI.

A WIRED and Indicator review found sexualized deepfake incidents at about 90 schools in at least 28 countries since 2023. The Internet Watch Foundation said Report Remove received 1,894 reports from under-18s in 2025, up 66%. A teen in New Jersey and another in Ohio were charged over AI nudes of classmates, and an Ohio man became the first person convicted under the Take It Down Act. NPR reported that such cases are getting harder to prosecute. The allegations are unproven.

What to watch

  • Ofcom’s review of the six platforms’ responses to its 30 April deadline.
  • The second phase of New Mexico’s case against Meta, which begins in May.
  • Whether the EU restores voluntary scanning rules.
  • The UK consultation, which closes on 26 May.
  • Turkey’s presidential approval of the under-15 bill, and Greece’s January 2027 start.
Sources (57)
  1. EU interim ePrivacy derogation for voluntary CSAM detection expires digitalpolicyalert.org
  2. Greece announces social media ban for children under 15 from January 2027 government.gov.gr
  3. Turkish parliament passes bill restricting social media for under-15s AP News
  4. Roblox on the hook for $12.5M in Nevada online child safety settlement biometricupdate.com
  5. Ban on kid companion chatbots advanced by Senate committee rollcall.com
  6. EU rules allowing platforms to scan for child sexual abuse material lapsed on April 3, Thorn warned Thorn
  7. EU lapse in child safety rules raises concerns blog.google
  8. European Commission president announced the rollout of the EU digital age verification app European Commission
  9. European Commission urges fast rollout of EU age verification app digital-strategy.ec.europa.eu
  10. Meta faces EU Digital Services Act breach finding over under-13 access ec.europa.eu
  11. Turkish lawmakers passed a bill restricting social media use for children under 15 Al Jazeera
  12. Macron pushed for a blanket social media ban for under-15s Reuters
  13. MPs rejected an under-16 social media ban for a second time The Guardian
  14. UK House of Commons backs amendments in lieu on Children’s Wellbeing Bill with online safety provisions hansard.parliament.uk
  15. Starmer told social media bosses ’things can’t go on like this’ as the UK weighs an under-16 ban BBC News
  16. UK ministers table amendment that could delay under-16 social media limits by three years The Guardian
  17. UK online safety consultation drew nearly 50,000 responses, about 6,000 from children The Independent
  18. UK proposes jail for tech executives whose platforms ignore Ofcom intimate image rulings Reuters
  19. National Crime Agency to receive CSEA reports under UK Online Safety Act rules legislation.gov.uk
  20. Ofcom opened investigations into Telegram over child sexual abuse material and into two teen chat sites Ofcom
  21. Meta must face Massachusetts lawsuit over youth social media addiction, court rules theguardian.com
  22. Meta threatens to shut down social networks in New Mexico over child safety court case theguardian.com
  23. West Virginia announced an $11 million settlement with Roblox over child safety WV MetroNews
  24. Alabama secures $12.2 million Roblox settlement, mandates sweeping child safety reforms gulfcoastmedia.com
  25. Man jailed after grooming a 14-year-old he met on Roblox BBC News
  26. Florida attorney general probes OpenAI over alleged risks to minors wogx.com
  27. Uthmeier expands criminal AI investigation to USF student murders orlandoweekly.com
  28. Families of seven Tumbler Ridge shooting victims sued OpenAI over failure to report the shooter The Guardian
  29. Grassley opened an inquiry into eight tech firms over child exploitation reporting Office of Senator Chuck Grassley
  30. Mississippi governor signed the Keeping Kids Safe Online Act Magnolia Tribune
  31. Kansas Governor signs Caleb’s Law, targeting online sextortion of minors kctv5.com
  32. Wisconsin governor signed five bills on child internet safety WMTV
  33. Nebraska passes bill adding safeguards for minors using conversational AI Nebraska Legislature
  34. Michigan Senate passed a bill regulating companion chatbots for minors DataGuidance
  35. California Assembly committee advances under-16 social media ban ABC10 Sacramento
  36. Reps. Foushee, Moore Introduce Bipartisan Bill Protecting Children from AI Companion Chatbots Congresswoman Valerie Foushee (.gov)
  37. Australia backs Digital Duty of Care after Online Safety Act review International Justice Mission (via News Hub)
  38. Survey finds 70% of parents say their child still uses social apps despite Australia’s ban socialmediatoday.com
  39. Australian teens say social media ban is futile and look to masks, parent ID to dodge controls fortune.com
  40. Indonesia sent YouTube a letter of reprimand over children’s social media rules Reuters
  41. New rules on AI set to protect minors China Daily
  42. China proposes rules on digital humans and bans intimate virtual relationships for minors Reuters
  43. Canadian Liberal members voted to back age limits on social media and AI chatbots CBC News
  44. Manitoba Moves Toward Social Media Ban For Kids as Canada Eyes AI chatbot Use Restriction techtimes.com
  45. Roblox introduced two age-specific account types with expanded checks, and parents raised concerns BBC News
  46. Roblox adds youth accounts as push for social media bans grows eastbaytimes.com
  47. Instagram expanded 13+ content ratings for teen accounts worldwide Meta Newsroom
  48. Meta expands parental oversight with new AI conversation insights for teens about.fb.com
  49. OpenAI introduced a Child Safety Blueprint on AI-enabled child sexual exploitation openai.com
  50. Common Sense Media asked major AI companies to give $100 million each for independent child-impact research Politico
  51. Drexel study found teens described addiction-like reliance on Character.AI Drexel University
  52. WIRED review found sexual deepfake abuse at about 90 schools worldwide WIRED
  53. IWF reports UK child sextortion reports surge as Report Remove use rises iwf.org.uk
  54. New Jersey teen charged over AI-generated nude images of at least 32 girls The Montgomery News
  55. Second Mason High student charged over AI nudes of classmates Fox19 / Cincinnati Enquirer
  56. First conviction under the Take It Down Act follows guilty plea in Ohio Ars Technica
  57. NPR reported AI-generated abuse imagery cases were getting harder to prosecute NPR

Every story recorded for April 2026 is in the news archive. You can also search all stories.

John Zoltner is the founder and CEO of AIChildSafety.org. About John