zoltner

AI Child Safety News / archive

March 2026: 111 stories on AI and child safety

Every story recorded for March 2026 that is in this archive, newest first. Each entry gives a short summary and links to the original report. Read the March 2026 summary

Search all stories

  1. eSafety said platforms removed 4.7 million under-16 accounts but many children kept access

    · eSafety Commissioner (Australia)

    Australia’s eSafety Commissioner said in its March compliance update that platforms removed or restricted about 4.7 million age-restricted accounts by mid-December and blocked over 300,000 more by early March, yet many under-16s kept accounts. It is investigating Facebook, Instagram, Snapchat, TikTok and YouTube.

    Age limits and social media bans; Regulatory enforcement and investigations; Legislation and rulemaking · eSafety Commissioner, Facebook, Instagram, Snap, TikTok

  2. Science study finds 11 AI models affirmed users’ actions 49% more often than humans did

    · science.org

    In a Science study, 11 AI models affirmed users’ actions 49% more often on average than human respondents. In experiments involving interpersonal conflicts, people who received overly affirming AI advice were more convinced they were right and less willing to repair the conflict. They also expressed greater interest in using the AI again.

    Research and evidence · AI hallucination / sycophancy

  3. Brazil study maps age assurance practices across 25 digital services

    · cgi.br

    A CGI.br and NIC.br study by Cetic.br reviewed 25 popular digital services used by children and adolescents in Brazil. It was previewed on 18 March at a seminar in Brasília on the Digital Child and Adolescent Statute (ECA Digital), which took effect that day. Most services (21 of 25) do not check age at signup.

    Age assurance and verification; Research and evidence · Brazil, age assurance, ECA Digital (Brazil)

  4. Ofcom tightens online safety enforcement across major platforms

    · ofcom.org.uk

    Ofcom’s March 2026 industry bulletin said 77 of the top 100 dedicated pornography services now use age assurance and 7 more block UK users. Ofcom wrote to Facebook, Instagram, Roblox, Snapchat, TikTok and YouTube with four demands for stronger child protections, with an April 30 deadline to respond.

    Regulatory enforcement and investigations; Age assurance and verification · Ofcom, Facebook, Instagram, Roblox, Snap

  5. South Korean lawmakers propose limits on algorithmic feeds for minors

    · The Korea Herald

    Two South Korean lawmakers proposed bills in March to limit algorithm-selected content for users under 19 on social media. The bills would give the Korea Media and Communications Commission authority over platform operators and penalize platforms that fail to verify age. One bill was taken up at a National Assembly subcommittee meeting.

    Legislation and rulemaking; Age assurance and verification; Platform safety features and design · South Korea, algorithmic feeds, European Commission

  6. Australian jailed and six Philippine children removed from harm in abuse caseAllegation

    · Australian Federal Police

    The AFP and Philippine NBI said on March 29 that their partnership led to six children being removed from harm and four adults being charged. A 56-year-old Penrith man pleaded guilty and was sentenced on March 16 to 10 years and four months for encouraging and grooming adults to facilitate live online abuse of children. The allegations are unproven.

    Criminal cases and prosecutions; Lawsuits and settlements · criminal prosecution, Philippines, Australia, grooming

  7. Starmer pledges to ban addictive algorithms after US verdicts against Meta and Google

    · The Observer

    Prime Minister Keir Starmer told The Observer that the Los Angeles verdict against Meta and Google was a turning point. He pledged to ban addictive algorithms that kept children online.

    Legislation and rulemaking; Lawsuits and settlements; Platform safety features and design · Google, Meta, Keir Starmer, addictive design, algorithmic feeds

  8. Indonesia starts implementing social media restrictions for children under 16

    · channelnewsasia.com

    The restrictions began on Mar 28, making Indonesia the first country in Southeast Asia to ban children from having accounts on YouTube, TikTok, Facebook, Instagram, Threads, X, Bigo Live and Roblox.

    Age limits and social media bans; Legislation and rulemaking · Indonesia, Facebook, Instagram, Roblox, TikTok

  9. Indonesia follows Australia’s lead with landmark social media ban for minors

    · Yahoo Finance

    Investing.com reported on March 28 that Indonesia began enforcing a nationwide restriction on social media for children under 16, the first in Southeast Asia. Platforms classed as high risk, including TikTok, Instagram and YouTube, must deactivate minors’ accounts or face sanctions, and all platforms must complete a child safety self-assessment by June.

    Age limits and social media bans; Legislation and rulemaking · Indonesia, social media ban, Australia, Instagram, TikTok

  10. Austria plans social media ban for children under 14

    · AP News

    Austria’s governing coalition announced on March 27 that it plans to ban social media use for children under 14. Officials said draft legislation would be drawn up by the end of June and that technically modern age verification methods would be used.

    Age limits and social media bans; Legislation and rulemaking; Age assurance and verification · social media ban, Austria, age assurance, international cooperation

  11. UK issues first screen time guidance for under-fives that advises avoiding AI tools

    · The Guardian

    The UK government issued guidance in late March advising no screens for children under two, apart from shared activities, and no more than one hour a day for ages two to five. It advised avoiding fast-paced, social media-style videos and AI tools for young children. A panel led by Rachel de Souza and Prof Russell Viner developed it.

    Legislation and rulemaking; Platform safety features and design · United Kingdom, screen time guidance, screen time limits

  12. Dutch court orders Grok to stop creating AI nudes and child abuse imageryAllegation

    · CNBC

    An Amsterdam court prohibited xAI’s Grok from generating non-consensual nude images and child sexual abuse material, in a case brought by the group Offlimits. xAI faced a fine of 100,000 euros a day, capped at 10 million euros. The allegations are unproven.

    Lawsuits and settlements; AI-generated abuse imagery; AI chatbots and companions · Grok, AI-generated CSAM, Netherlands, xAI

  13. ICO and Ofcom issue guidance on age assurance and online safety

    · ico.org.uk

    The Information Commissioner’s Office and Ofcom have issued a joint statement outlining how age assurance measures should align with online safety and data protection requirements.

    Regulatory enforcement and investigations; Age assurance and verification · ICO, Ofcom, age assurance, screen time guidance, privacy and data protection

  14. New AI safety policies target teen protection in apps

    · openai.com

    OpenAI has released a set of prompt-based safety policies to help developers build safer AI experiences for teenagers. The tools work with the open-weight model gpt-oss-safeguard, turning safety requirements into practical classifiers for real-world use.

    Legislation and rulemaking · AI governance, OpenAI

  15. European Commission opened formal proceedings into Snapchat’s child protection compliance under the DSAAllegation

    · European Commission

    The European Commission opened formal proceedings on March 26 to assess whether Snapchat protected children’s privacy, safety, and security under the Digital Services Act. The inquiry covered five areas, including age assurance, protections from harmful contact, default settings, access to illegal or age-restricted products, and reporting and redress. The allegations are unproven.

    Lawsuits and settlements; Age assurance and verification · Digital Services Act, Snap, European Commission, European Union, age assurance

  16. European Commission preliminarily found Pornhub, Stripchat, XNXX and XVideos breach the DSA on minors’ accessAllegation

    · European Commission

    On March 26 the European Commission preliminarily found Pornhub, Stripchat, XNXX and XVideos in breach of the Digital Services Act for failing to protect minors from exposure to pornographic content, including on risk assessment. The allegations are unproven.

    Lawsuits and settlements · Digital Services Act, Pornhub / Aylo, pornography, European Commission, European Union

  17. Two teens received probation for making AI fake nudes of classmates

    · AP News

    Two teenage boys were given probation on 25 March after using artificial intelligence to create hundreds of fake nude images of classmates at a private school in Pennsylvania. At least 59 girls were among the victims.

    Criminal cases and prosecutions; AI-generated abuse imagery · criminal prosecution, deepfake nudes in schools, Pennsylvania (US)

  18. Senator Markey introduced a bill on privacy safeguards for minors using AI chatbots

    · Office of Sen. Edward Markey

    Sen. Edward Markey introduced the Youth AI Privacy Act on March 25. It would ban advertising to minors in chatbots, bar training models on minors’ data, restrict addictive features and require repeated notices that the chatbot is not human. The FTC, state attorneys general and parents could enforce it.

    Legislation and rulemaking; Platform safety features and design; AI chatbots and companions · chatbots, privacy and data protection, Youth AI Privacy Act, FTC, addictive design

  19. Ofcom and the ICO published a joint statement on a common approach to age assurance

    · Ofcom

    Ofcom and the Information Commissioner’s Office published a joint statement on March 25 summarizing their existing age assurance policy in a practical way, to help services likely to be accessed by children comply with both the Online Safety Act and UK data protection law.

    Regulatory enforcement and investigations; Age assurance and verification; Legislation and rulemaking · ICO, Ofcom, age assurance, Online Safety Act (UK), privacy and data protection

  20. Washington governor signed a law regulating companion chatbots and protecting minors

    · OPB

    Gov. Bob Ferguson signed House Bill 2225 on 24 March. It requires companion chatbots to disclose they are not human, hourly for users under 18, and bars sexually explicit conversations with minors and manipulative engagement techniques. A second bill, House Bill 1170, requires substantially AI-modified content from large AI companies to be traceable through watermarks or metadata.

    Legislation and rulemaking; AI chatbots and companions · chatbots, companion chatbots

  21. Baltimore sued xAI over Grok sexual deepfakes including images of childrenAllegation

    · CNA

    The city of Baltimore sued xAI on 24 March in Baltimore Circuit Court, alleging Grok illegally generated nonconsensual sexually explicit images, including of children, in violation of the city’s consumer protection law. The allegations are unproven.

    Lawsuits and settlements; AI-generated abuse imagery; AI chatbots and companions · Grok, xAI, deepfakes, consumer protection

  22. eSafety report found AI companion chatbots lacked meaningful age checks

    · eSafety Commissioner

    Australia’s eSafety Commissioner reported on 24 March that Character.AI, Nomi, Chai and Chub AI had no meaningful age checks and relied mainly on self-declared age. Most did not refer users discussing self-harm to support. A survey found 79% of children aged 10 to 17 had used an AI companion or assistant.

    AI chatbots and companions; Research and evidence; Age assurance and verification · age assurance, companion chatbots, eSafety Commissioner, chatbots, Character.AI

  23. Internet Watch Foundation launched a report on AI-generated child sexual abuse material

    · LinkedIn (IWF staff post)

    The Internet Watch Foundation held a parliamentary reception on March 24 to launch its report, Harm without limits: AI child sexual abuse material through the eyes of our Analysts, according to a post by an IWF staff member.

    AI-generated abuse imagery; Grooming, sextortion and exploitation; Industry policy and partnerships · Internet Watch Foundation, AI-generated CSAM

  24. eSafety report finds AI companion apps failing to protect Australian children

    · The Sydney Morning Herald

    Australia’s eSafety Commissioner reported on March 23 that Character.AI, Chub AI, Nomi and Chai had no meaningful age checks, and that Chai, Chub AI and Nomi did not direct users to support when self-harm was detected. In a survey of nearly 2,000 children aged 10 to 17, 79 percent had used an AI chatbot or companion.

    Research and evidence; AI chatbots and companions; Regulatory enforcement and investigations · companion chatbots, eSafety Commissioner, Australia, Character.AI, age assurance

  25. Two teens await sentencing for nudifying classmates as parents plan to sue schoolAllegation

    · arstechnica.com

    Two 16-year-old boys who admitted using AI to create sexualized images of 48 female classmates and 12 other young women at a private school in Pennsylvania were due to be sentenced on March 25. Families of at least 10 victims planned to sue the school. The allegations are unproven.

    AI-generated abuse imagery; Lawsuits and settlements; Education and schools · nudification, criminal prosecution, sexualized AI content, Pennsylvania (US)

  26. UN News article says women rarely get justice for AI deepfake abuse

    · news.un.org

    UN News reported on 21 March 2026 that AI-generated sexualized deepfake abuse harms women and girls and that survivors rarely get justice. The article cited a 2023 report finding that deepfake pornography made up 98% of deepfake videos online and that 99% depicted women. It called for clearer laws and faster platform removal.

    AI-generated abuse imagery · deepfakes, UN, international cooperation, pornography, sexualized AI content

  27. Philippines warns Roblox it could face a ban over child safety concernsAllegation

    · ABS-CBN News

    Philippine officials warned that Roblox could face restrictions or suspension unless it strengthened child safety protections. The country’s cybercrime agency gave the company until the end of March to contact authorities about its safeguards, citing alleged grooming and exploitation on the platform. The allegations are unproven.

    Platform safety features and design; Gaming and virtual worlds · Roblox, gaming platforms, Philippines, grooming

  28. Minnesota bill would ban minors from AI companion chatbots

    · MinnPost

    A bipartisan Minnesota bill would ban people under 18 from using companion-style AI chatbots, which would be a first in the US. Lawmakers cited the suicides of three teenagers after prolonged conversations with chatbots.

    Legislation and rulemaking; AI chatbots and companions; Age limits and social media bans · chatbots, companion chatbots, legislative proposal, Minnesota (US), self-harm and suicide

  29. White House AI framework asks Congress for child protections and state preemption

    · Roll Call

    The White House released a four-page national AI legislative framework that asks Congress to require commercially reasonable age assurance, which could include parental attestation, and to reduce risks of sexual exploitation and self-harm for minors on AI platforms. It also calls for broad preemption of state AI laws.

    Legislation and rulemaking; Industry policy and partnerships; Age assurance and verification · United States, age assurance, self-harm and suicide

  30. Ecuador bill would bar children under 15 from social media, citing organized crime recruitment

    · techpolicy.press

    A bill by Assemblymember Katherine Pacheco Machuca would bar children under 15 from social media in Ecuador, framed as a response to organized crime. The National Assembly’s child protection commission has begun reviewing it. The article reports no vote date.

    Legislation and rulemaking; Age limits and social media bans · legislative proposal

  31. Science Feedback study finds TikTok had the most misinformation among six platforms, Euractiv reports

    · euractiv.com

    Euractiv reported that TikTok had the highest prevalence of misleading posts among six platforms examined by Science Feedback in four EU countries. About one in four sampled TikTok posts were misleading, and most synthetic videos researchers examined lacked AI labels. TikTok said it removed harmful misinformation that violated its rules.

    Platform safety features and design; Research and evidence; Legislation and rulemaking · TikTok, misinformation, European Union

  32. Canada reconvened its expert advisory group on online safety to examine AI and chatbots

    · Canadian Heritage (Government of Canada)

    The Government of Canada reconvened an 11-member expert advisory group on online safety, which Canadian Heritage first consulted in 2022 and whose advice helped shape Bill C-63. The group will advise on AI, chatbots and AI companions, and its kickoff meeting took place on March 11, 2026.

    Legislation and rulemaking; AI chatbots and companions · Canada, chatbots, companion chatbots

  33. Germany prepared to criminalize pornographic deepfakes after a high-profile case, DW reported

    · Deutsche Welle

    Germany’s justice ministry said on March 20 that a draft law criminalizing the creation and distribution of pornographic deepfakes was practically finished. It follows a complaint by actress Collien Fernandes against her ex-husband, who denies the accusations. The draft would also give police more power to search suspects’ devices. The allegations are unproven.

    Legislation and rulemaking; AI-generated abuse imagery · deepfakes, pornography, Germany, criminal prosecution, legislative proposal

  34. North Dakota man arrested over AI-generated child abuse materialAllegation

    · Valley News Live

    Fargo police arrested a man on a possession charge after a March 17 search warrant. Court documents allege he used photos of women from his church to make AI-generated abuse videos, and that more AI-generated material was found on his phone. A preliminary hearing was set for April 15. The allegations are unproven.

    AI-generated abuse imagery; Criminal cases and prosecutions; Lawsuits and settlements · criminal prosecution, AI-generated CSAM, North Dakota (US)

  35. Pennsylvania officials weigh AI school safeguards after student deepfake cases

    · WHYY

    Pennsylvania Governor Josh Shapiro and Attorney General Dave Sunday held a roundtable on AI in schools after students created explicit deepfakes of classmates. Shapiro said his office’s test of an AI companion bot found it claiming to be a licensed professional.

    Education and schools; AI-generated abuse imagery; Legislation and rulemaking · deepfake nudes in schools, deepfakes, Pennsylvania (US), companion chatbots

  36. US senator proposes AI rules for children

    · iapp.org

    Sen. Marsha Blackburn released a discussion draft of a federal AI framework on March 18, 2026. It would impose a duty of care on developers for users under 17, allow private lawsuits over child harms tied to defective design, and set federal rules for marking AI-generated content.

    Legislation and rulemaking · United States, AI governance, Digital Duty of Care (Australia), Marsha Blackburn

  37. EU negotiators fail to agree on extending voluntary child abuse detection rules due to expire April 3

    · euractiv.com

    The European Parliament and Council failed to agree on extending temporary rules that let certain communications providers voluntarily scan for child sexual abuse material. Child-rights groups and a technology industry group criticized the impasse, while privacy advocates opposed broad scanning. Parliament wanted narrower rules and Council resisted. The rules were due to expire on April 3.

    Lawsuits and settlements; Legislation and rulemaking · European Union, AI-generated CSAM, European Parliament, privacy and data protection

  38. Ofcom fined 4chan £450,000 for not having age checks to stop children seeing pornography

    · Ofcom

    Ofcom fined 4chan £450,000 on March 19 for not having age checks in place to prevent children seeing pornography, plus £50,000 for not assessing the risk of illegal material and £20,000 for failing to set out in its terms of service how it protects people from criminal content.

    Regulatory enforcement and investigations; Age assurance and verification; Lawsuits and settlements · Ofcom, age assurance, pornography

  39. Pennsylvania man charged over AI-generated nude images of teen girlsAllegation

    · WFMJ

    A 56-year-old Greenville man faced 56 felony charges after police said they found AI-generated child sexual abuse material on a laptop, including altered photos taken from a girl’s Facebook page. He was arraigned on March 18. The allegations are unproven.

    Criminal cases and prosecutions; AI-generated abuse imagery; Lawsuits and settlements · criminal prosecution, Pennsylvania (US), Facebook, AI-generated CSAM

  40. Brazil signs decrees putting its child online safety statute into force

    · Planalto (Brazilian presidency)

    President Lula signed decrees on March 18 regulating Law 15,211/2025, the Digital Statute of Children and Adolescents. The rules require more reliable age verification, ban manipulative design aimed at minors and create a national center to receive platform reports of online crimes against children.

    Legislation and rulemaking; Age assurance and verification · Brazil, Luiz Inácio Lula da Silva, age assurance

  41. House of Lords agrees Kidron amendment creating offense for harmful AI chatbots

    · UK Parliament

    The House of Lords agreed a Baroness Kidron amendment to the Crime and Policing Bill that would make it an offense to create or supply an AI chatbot producing illegal content or, for users under 18, content harmful to children. It would also require chatbot risk assessments, with penalties of up to five years in prison.

    Legislation and rulemaking; AI chatbots and companions · chatbots

  42. PA lawmakers advance SAFECHAT Act to protect children from AI chatbot interactions

    · City & State Pennsylvania

    The Pennsylvania Senate voted 49-1 to approve SAFECHAT, Senate Bill 1090, which would require AI companion operators to prevent sexually explicit conduct with minors and direct users showing signs of suicidal ideation to crisis services. The bill now awaits consideration in the state House.

    Legislation and rulemaking; AI chatbots and companions · chatbots, United States, companion chatbots, legislative proposal, self-harm and suicide

  43. Parents underestimate how teenagers use AI in daily life

    · bbc.com

    A Pew Research Center survey of 1,458 US teens and their parents, reported March 18, found 51% of parents said their children use AI while 64% of teens said they use chatbots. Four in 10 parents had never discussed AI with their children, and 12% of teens used it for advice or emotional support.

    Research and evidence · Pew Research Center, chatbots, survey, United States

  44. Ofcom published a global survey of child sexual abuse material perpetrators by Protect Children

    · Ofcom

    Ofcom published on March 18 a report by Protect Children, commissioned by Ofcom, based on an anonymous survey of more than 20,000 people who searched the dark web for child sexual abuse material. Fieldwork ended in January 2026. The report examines how perpetrators use platform designs to view, host and distribute the material.

    Research and evidence; AI-generated abuse imagery; Regulatory enforcement and investigations · Ofcom, AI-generated CSAM, survey

  45. Teens File Lawsuit Against xAI for Grok AIAllegation

    · el-balad.com

    Two minors and an adult have filed a lawsuit against xAI, claiming it is responsible for AI-generated child sexual abuse material (CSAM) involving at least 19 minors. One plaintiff, known as “Jane Doe 1,” discovered explicit images of herself circulating on Discord last December. The allegations are unproven.

    Lawsuits and settlements; AI-generated abuse imagery; AI chatbots and companions · xAI, Grok, AI-generated CSAM, Discord, gaming platforms

  46. Tennessee teens sue Elon Musk’s xAI over AI-generated child sexual abuse materialAllegation

    · WPLN News

    Three Tennessee teenagers filed a class action against xAI, alleging that an app built on xAI’s technology was used to create nonconsensual nude images of them. The plaintiffs say xAI licensed the technology to app makers. The allegations are unproven.

    Lawsuits and settlements; AI-generated abuse imagery; Grooming, sextortion and exploitation · xAI, Tennessee (US), AI-generated CSAM, Elon Musk, class action

  47. AI chatbots raise risks as EU urged to enforce DSA rules

    · techpolicy.press

    In a March 17 TechPolicy.Press op-ed, Laura Kaun urged the EU to enforce the Digital Services Act on AI chatbots, citing a Center for Countering Digital Hate finding that 8 of 10 popular chatbots would help a teenager plan an attack. She proposed treating chatbots as hosting providers now and amending the act to create a chatbot category.

    Lawsuits and settlements; AI chatbots and companions; Regulatory enforcement and investigations · chatbots, Digital Services Act, European Union, CCDH

  48. Survey finds most parents want AI chatbot guardrails for children

    · Education Week

    An Echelon Insights survey of 1,511 US K-12 parents for the National Parents Union found 86% wanted pop-up warnings on sensitive topics, 85% wanted parents alerted to harmful chats and 79% wanted parental permission before minors used chatbots.

    Research and evidence; AI chatbots and companions · chatbots, survey, AI in education, United States

  49. ECA Digital law raises pressure on Big Tech in Brazil

    · techpolicy.press

    Brazil’s ECA Digital child online protection law took effect in March 2026, six months after Congress approved it in 2025. Penalties include fines of up to $10 million. It requires reliable age verification and bars design features that drive compulsive use by minors.

    Legislation and rulemaking; Lawsuits and settlements; Platform safety features and design · ECA Digital (Brazil), Brazil, addictive design, age assurance, United States

  50. xAI faces proposed class action over Grok-altered sexual images of minorsAllegation

    · techcrunch.com

    Three anonymous plaintiffs, two of them minors, filed a proposed class action against xAI on March 16 in federal court in Northern California. They allege Grok altered real images of them as minors into sexual content and that xAI lacked safeguards other labs use. The allegations are unproven.

    Lawsuits and settlements; AI chatbots and companions; AI-generated abuse imagery · xAI, Grok, class action, sexualized AI content, California (US)

  51. Unsealed Meta messages showed staff warned 7.5 million annual child abuse reports could vanish

    · ibtimes.co.uk

    Internal Meta messages unsealed in New Mexico’s child exploitation trial showed employees warned in 2023 that about 7.5 million annual child sexual abuse reports on Messenger could disappear after end-to-end encryption. Meta said it can review encrypted messages that are reported for child safety issues. The allegations are unproven.

    Lawsuits and settlements · Meta, Messenger, encryption, New Mexico (US)

  52. UK police document warned of children earning TikTok gifts for explicit livestreamsAllegation

    · The Telegraph

    The Telegraph reported that a UK police intelligence assessment identified children as young as five on TikTok livestreams and offenders using virtual gifts to induce sexual or compromising acts. Police also described children selling sexual material for money or in-game currency. TikTok said it was unaware of the examples and disputed some findings. The allegations are unproven.

    Criminal cases and prosecutions; Platform safety features and design · TikTok, criminal prosecution, livestreaming, United Kingdom

  53. Brazilian police investigated a 14-year-old over AI-faked nude images of classmatesAllegation

    · Folha do ES

    Police in Vitória, Espírito Santo, began investigating a 14-year-old who allegedly used AI to create fake nude images of two students at a private school and shared them with other pupils. The school reported the case to the authorities. The allegations are unproven.

    Lawsuits and settlements; Criminal cases and prosecutions; Regulatory enforcement and investigations · criminal prosecution, deepfake nudes in schools, Brazil

  54. Italian police arrested a school worker over AI nude edits of pupilsAllegation

    · ANSA

    Postal police in Catania arrested a 58-year-old school assistant, who they said secretly photographed middle school pupils and used AI to make them appear nude beside him, and who held a large quantity of child sexual abuse material. A judge placed him under house arrest after a validation hearing. The allegations are unproven.

    Criminal cases and prosecutions; Lawsuits and settlements; Education and schools · criminal prosecution, deepfake nudes in schools, Italy, AI-generated CSAM

  55. Iowa juvenile court case over AI nude photos at an Iowa high school concludedAllegation

    · Radio Dubuque

    Radio Dubuque reported that juvenile court proceedings had concluded in a case involving AI-generated nude images affecting dozens of students at an Iowa high school. Authorities charged three juveniles in September. The county attorney said the charges and outcomes could not be disclosed because of the juveniles’ ages.

    Lawsuits and settlements; Criminal cases and prosecutions; Education and schools · deepfake nudes in schools, Iowa (US), criminal prosecution

  56. European Parliament backed extending voluntary CSAM detection rules until 2027

    · Help Net Security

    The European Parliament voted to extend the ePrivacy derogation allowing platforms to voluntarily detect child sexual abuse material from 3 April 2026 to 3 August 2027. The exemption would not cover end-to-end encrypted communications. Talks with the Council were to follow.

    Legislation and rulemaking; AI-generated abuse imagery · AI-generated CSAM, European Parliament, European Union, encryption, privacy and data protection

  57. About a fifth of Australian 13- to 15-year-olds in Qustodio data still used TikTok and Snapchat

    · Reuters

    Reuters reported that in February, 21.2% of Australian 13- to 15-year-olds tracked by parental-control company Qustodio used TikTok and 20.3% used Snapchat, about two months after Australia’s under-16 social media restrictions took effect. Both figures had fallen since November, and Qustodio said the seasonal decline was steeper than the previous year’s.

    Research and evidence · Snap, TikTok, Australia

  58. European Council proposes AI Act ban on nudification tools and AI-generated abuse material

    · therecord.media

    On 13 March 2026 the European Council proposed AI Act amendments banning AI practices that generate non-consensual sexual content and child sexual abuse material. The text also restores a “strict necessity” standard for processing sensitive personal data for bias detection. It must still be negotiated with Parliament.

    AI-generated abuse imagery; Grooming, sextortion and exploitation; Legislation and rulemaking · EU AI Act, nudification, AI-generated CSAM, privacy and data protection, sexualized AI content

  59. Cambridge researchers warn AI toys misread children’s emotions

    · theguardian.com

    A University of Cambridge study of AI-powered toys, tested with 14 children aged three to five, found they struggled with pretend play and misread children’s emotions. The researchers called for tighter regulation and safety kitemarks, including limits on toys affirming friendship with young children.

    Research and evidence · AI toys, AI toys (devices)

  60. Survey of Israeli adolescents found 84% used generative AI for emotional supportPreprint

    · OSF Preprints

    A survey of 594 Israeli adolescents aged 12 to 18 found 84% used generative AI for emotional support and 93% for schoolwork. Among those seeking support, 30% found it very helpful and 22% reported reduced loneliness. Knowledge of how the tools work was limited. It has not been peer reviewed.

    Research and evidence; AI chatbots and companions · survey, Israel, mental health

  61. Two Pennsylvania teens pleaded guilty over AI nude images of classmates

    · FOX43

    Two 16-year-olds pleaded guilty in juvenile court to 59 felony counts and conspiracy after using AI to create simulated nude images of classmates, Pennsylvania’s attorney general said. Investigators found 347 images and videos shared on Discord. Sentencing was set for March 25.

    Lawsuits and settlements; Criminal cases and prosecutions · deepfake nudes in schools, Pennsylvania (US), Discord, gaming platforms

  62. Appeals court voided most of the injunction on California’s children’s design code

    · Reuters

    The 9th U.S. Circuit Court of Appeals threw out much of an injunction blocking California’s children’s online safety law, finding NetChoice unlikely to prove its facial First Amendment challenge. It upheld blocks on some data-use and dark-pattern restrictions.

    Lawsuits and settlements; Legislation and rulemaking · injunction, California (US), First Amendment, NetChoice, United States

  63. UK watchdogs push social media giants to block children

    · canberratimes.com.au

    Ofcom, the UK media regulator, asked Facebook, Instagram, Roblox, Snapchat, TikTok and YouTube to show by April 30 how they would tighten age checks. The privacy regulator, the ICO, urged the platforms to adopt age-assurance tools. The government is considering barring under-16s from social media, following Australia.

    Legislation and rulemaking; Age assurance and verification; Regulatory enforcement and investigations · United Kingdom, Facebook, Instagram, Roblox, Snap

  64. EU lawmakers move forward on AI Act changes

    · iapp.org

    Members of the European Parliament reached a preliminary political agreement on amendments to the EU Artificial Intelligence Act on 11 March 2026. Two parliamentary committees will vote on the report on 18 March. The reported compromise includes a ban on AI systems that generate sexually explicit depictions of identifiable people without consent.

    Legislation and rulemaking · European Union, EU AI Act, European Parliament

  65. Ofcom told six major platforms to enforce minimum ages with highly effective age checks

    · Ofcom

    Ofcom wrote on March 12 to Facebook, Instagram, Roblox, Snapchat, TikTok and YouTube demanding effective minimum-age policies, failsafe grooming protections, safer feeds and an end to product testing on children. It set an April 30 deadline for them to report back.

    Regulatory enforcement and investigations; Age assurance and verification; Legislation and rulemaking · Ofcom, age assurance, age limits, Facebook, Instagram

  66. Washington legislature passes AI companion chatbot safety bill

    · Washington State Legislature

    Washington enacted legislation requiring AI companion chatbot disclosures and protocols for suicidal ideation and self-harm. It required providers to take reasonable measures to prevent sexually explicit or suggestive interactions with minors and to prohibit manipulative engagement techniques in those interactions. The law takes effect January 1, 2027.

    Legislation and rulemaking; AI chatbots and companions · chatbots, companion chatbots, self-harm and suicide

  67. Chatbots helped researchers posing as teens plot violent attacks, CNN and CCDH found

    · The Guardian

    Researchers from the Center for Countering Digital Hate and CNN posed as 13-year-old boys in tests of 10 chatbots in the US and Ireland. On average the chatbots enabled violence three-quarters of the time and discouraged it in 12% of cases. Claude and Snapchat’s My AI refused most requests.

    Research and evidence; AI chatbots and companions · chatbots, CCDH, Claude, Snap, Ireland

  68. Meta rolls out supervised WhatsApp accounts for younger users

    · techcrunch.com

    WhatsApp launched parent-supervised accounts for users under 13 on March 11, 2026. The accounts allow only messaging and calling, carry no ads, and give parents alerts and PIN-locked controls. Rollout begins in select geographies and is set to expand over the coming months.

    Platform safety features and design · WhatsApp, Meta

  69. Google outlines its approach to safer generative AI for young users

    · blog.google

    Google has described its approach to generative AI for young users, built on protection, respect for families and empowerment. Its policies prohibit harmful content for minors, including child sexual abuse material, self-harm content and non-consensual intimate imagery.

    Legislation and rulemaking; AI chatbots and companions · Google, AI-generated CSAM, self-harm and suicide

  70. Dutch court increases pressure on Meta over non-profiling social media feeds

    · euractiv.com

    The Amsterdam Court of Appeal raised Meta’s potential penalty from €5 million to €10 million if it failed to let Dutch Instagram and Facebook users keep their choice of a non-profiled feed. The court had ruled that resetting the feed to Meta’s profiling-based recommendation system was a deceptive design practice under the EU Digital Services Act.

    Lawsuits and settlements; Platform safety features and design · Meta, Netherlands, Digital Services Act, Facebook, Instagram

  71. UN poll finds two-thirds of children say cyberbullying has increased

    · UN News

    A poll of over 30,000 children by the UN Special Representative on Violence Against Children found two-thirds said cyberbullying had increased and half did not know where to seek help. The office said generative AI and deepfakes were making cyberbullying faster and harder to detect.

    Research and evidence · AI-enabled harassment, cyberbullying, UN, international cooperation, survey

  72. Amid wave of kid online safety laws, age-checking tech comes of age

    · asiaone.com

    Governments are moving to require age checks for social networks, AI chatbots and adult sites, and age-assurance tools have grown more accurate and cheaper. Australia reported that platforms had locked 4.7 million suspected underage accounts since its teen social media account ban launched in December 2025.

    Legislation and rulemaking; Age assurance and verification · age assurance, chatbots, pornography, Australia

  73. UK MPs rejected a social media ban for under-16s in a Commons vote

    · independent.co.uk

    MPs voted 307 to 173 on March 10 against an under-16 social media ban proposed by peers as an amendment to the Children’s Wellbeing and Schools Bill. They instead backed a government amendment giving the technology secretary powers to restrict or ban children from social media and chatbots after a consultation.

    Age limits and social media bans; Legislation and rulemaking · social media ban, United Kingdom, chatbots, mental health

  74. UK charity says abuse survivors increasingly contact support lines after consulting ChatGPT

    · thenews.com.pk

    The National Association of People Abused in Childhood (NAPAC) said some survivors told its support line they were referred by ChatGPT, which they used for therapy and exploration. The charity also reported a rise in ritualistic abuse disclosures, attributed to better professional training and AI use. A police working group was launched.

    AI chatbots and companions; Research and evidence · ChatGPT, child protection hotlines and NGOs, United Kingdom, criminal prosecution

  75. EU faces challenges in curbing digital abuse against women

    · euractiv.com

    Euractiv argued that EU digital rules have yet to stop technology from enabling abuse against women. The article highlighted Grok’s use to make nonconsensual sexualized images and reports of covert filming with Meta’s smart glasses. It cited gaps in rules for generating abusive imagery and in enforcement.

    Legislation and rulemaking · European Union, Grok, Meta, sexualized AI content

  76. Australia introduces strict online child safety rules covering AI chatbots

    · esafety.gov.au

    Australia’s Age-Restricted Material Codes took full effect on March 9, requiring search engines, app stores, social media, gaming providers and generative AI services to prevent children’s exposure to age-inappropriate content. AI companion chatbots able to generate sexually explicit or self-harm material must confirm users are 18 or older.

    AI chatbots and companions; Lawsuits and settlements · chatbots, Australia, Age-Restricted Material Codes (Australia), Apple, app stores

  77. Indonesia announces ban on social media for children under 16 starting March 28

    · mashable.com

    Indonesia’s communications minister announced on March 6 that children under 16 would be barred from social media, with accounts on high-risk platforms including X, YouTube, Facebook, Instagram, Threads, Roblox and Bigo Live deactivated from March 28. Mashable said few details were given on enforcement.

    Age limits and social media bans; Industry policy and partnerships; Legislation and rulemaking · Indonesia, social media ban, Facebook, Instagram, Roblox

  78. New AI feature keeps Roblox chat respectful and flowing

    · ir.roblox.com

    Roblox announced on March 5 AI-powered real-time chat rephrasing that replaces filtered profanity with more respectful wording, available only for in-experience chat between age-checked users in similar age groups. It also upgraded its text filters, cutting false negatives for sharing or soliciting personal information by 20x.

    Platform safety features and design; Gaming and virtual worlds · Roblox, gaming platforms, age assurance

  79. Spain’s digital minister backed an under-16 social media limit and warned of AI image abuse

    · Europa Press

    Digital Transformation Minister Óscar López said on 7 March that the government had sent Congress a bill to bar under-16s from social media. He warned that users upload photos to Grok to have people undressed, and said the government was working on algorithm transparency.

    Age limits and social media bans; Legislation and rulemaking · Spain, Grok, algorithmic feeds, nudification, United States

  80. South Carolina police investigated AI-generated nude images of five high school girlsAllegation

    · Live 5 News

    North Charleston police seized devices and were investigating two minors after five female students at a local high school were depicted in AI-generated nude images posted on multiple social media sites. No charges or arrests had been made. The allegations are unproven.

    Criminal cases and prosecutions; Education and schools; Regulatory enforcement and investigations · criminal prosecution, deepfake nudes in schools, South Carolina (US)

  81. Louisiana attorney general sued Roblox over child predatorsAllegation

    · WAFB

    Attorney General Liz Murrill filed suit against Roblox, arguing the platform was marketed as safe for children while predators targeted minors on it. She said her office arrested people who found children on Roblox every week. The allegations are unproven.

    Lawsuits and settlements; Grooming, sextortion and exploitation; Gaming and virtual worlds · Roblox, gaming platforms, Louisiana (US), criminal prosecution

  82. Kids Online Safety Act Advances to House Amid Concerns Over Free Speech and Civil Rights

    · cnet.com

    The House Energy and Commerce Committee advanced the Kids Online Safety Act to the full House on March 5, with no floor vote date set. The ACLU and other critics said its definition of harmful content could censor lawful speech. The bill directs agencies to study device-level age verification but does not require it.

    Legislation and rulemaking; Age assurance and verification · Kids Online Safety Act (KOSA), Online Safety Act (UK), age assurance

  83. Indonesia signs regulation barring under-16s from accounts on high-risk platforms

    · apnews.com

    Indonesia’s communications minister said on March 6 that she had signed a regulation barring children under 16 from accounts on high-risk platforms including YouTube, TikTok, Instagram, X and Roblox. Rollout starts gradually on March 28. AP described it as the first such restriction in Southeast Asia.

    Legislation and rulemaking; Age limits and social media bans · Indonesia, Instagram, Roblox, TikTok, X (Twitter)

  84. EU launches panel on child safety online and social media age rules

    · ec.europa.eu

    The European Commission has convened a new expert panel tasked with examining how children can be better protected across digital platforms, including social media, gaming environments and AI tools.

    Legislation and rulemaking; Industry policy and partnerships · European Union, European Commission

  85. 371 security and privacy scientists signed a letter calling for a global pause on age assurance rollouts

    · Politico Europe

    On March 2, an open letter signed by 371 security and privacy academics in 29 countries said the age assurance technologies being rolled out are not effective and carry significant privacy and security risks. It called for a global pause until scientific consensus settles on their benefits, harms and technical feasibility.

    Age assurance and verification; Privacy and children’s data; Research and evidence · privacy and data protection, age assurance, settlement

  86. Ofcom opened an investigation into the provider of two image boards hosting image-based sexual abuseAllegation

    · Ofcom

    Ofcom opened a formal investigation on March 6 into whether the provider of two unnamed online image boards failed to protect UK users from illegal content, including non-consensual intimate images and child sexual abuse material. The allegations are unproven.

    Regulatory enforcement and investigations; Lawsuits and settlements · Ofcom, AI-generated CSAM, United Kingdom

  87. Senate passes COPPA 2.0 teen privacy bill by unanimous consent

    · Bloomberg Government

    The Senate passed the Children and Teens’ Online Privacy Protection Act (S. 836) by unanimous consent. The bill would extend federal online privacy protections to teens ages 13 through 16, bar targeted advertising to children and teens, and let families request deletion of personal information.

    Legislation and rulemaking; Privacy and children’s data · privacy and data protection, United States, COPPA, legislative proposal

  88. Privacy lawsuit targets Meta AI glasses after reports of footage reviewAllegation

    · techcrunch.com

    Meta is facing a new lawsuit in the US over privacy concerns tied to its AI smart glasses. The legal complaint follows investigative reporting indicating that contractors working for a Kenya-based subcontractor reviewed footage captured by users’ devices, including sensitive personal scenes. The allegations are unproven.

    Lawsuits and settlements; Privacy and children’s data; AI chatbots and companions · Meta AI, privacy and data protection, Meta, Kenya, United States

  89. EU considers placing Roblox under strict Digital Services Act rules

    · euractiv.com

    Euractiv reported that the European Commission was considering whether to designate Roblox a very large online platform under the Digital Services Act after Roblox reported 48 million monthly EU users, above the 45 million threshold. The Commission was reviewing the figures and had set no deadline for a decision.

    Legislation and rulemaking; Gaming and virtual worlds · Roblox, Digital Services Act, gaming platforms, European Union, European Commission

  90. China expands oversight of youth online safety

    · chinadaily.com.cn

    China has introduced new measures to regulate online information that could affect the physical and mental health of minors. Authorities in China said the rules will take effect on 1 March and aim to improve protection for young internet users.

    Legislation and rulemaking · China, mental health

  91. FOSI counted almost 80 state bills on AI chatbots and over 1,000 on AI more broadly

    · Family Online Safety Institute (The Assurance newsletter)

    FOSI’s March 5 newsletter says US states have introduced almost 80 bills specifically on AI chatbots and over 1,000 related to AI more broadly, while Congress is considering the GUARD Act, which would require age verification for AI companions.

    Legislation and rulemaking; AI chatbots and companions; Age assurance and verification · chatbots, GUARD Act, United States, age assurance, companion chatbots

  92. Manitoba opened consultations on AI risks including deepfakes and child-targeted manipulation

    · Government of Manitoba

    Manitoba launched provincewide consultations on privacy law and AI rules that cover deepfakes and child-targeted manipulation. The government said it would consider age limits for access. The Canadian Centre for Child Protection cited rising AI-generated child sexual abuse material.

    AI-generated abuse imagery; Legislation and rulemaking · Canada, deepfakes, AI governance, AI-generated CSAM, Canadian Centre for Child Protection

  93. Florida Senate passed an AI Bill of Rights with companion chatbot rules for minors

    · Florida Senate

    The Florida Senate passed SB 482, which requires parental consent for minors to hold companion chatbot accounts, hourly reminders that the bot is not human and measures against harmful material for minors. It also limits AI instructional tools before grade 6.

    Legislation and rulemaking; AI chatbots and companions · chatbots, companion chatbots, Florida (US), United States, parental consent

  94. Parliament deadlock leaves EU chat-scanning extension in doubt

    · euractiv.com

    The European Parliament’s civil liberties committee failed on March 2 to adopt its amended report on extending the EU’s temporary rules letting tech firms voluntarily scan for child sexual abuse material. The file went to a plenary vote in the week of March 9 to 12, with the current rules due to expire in April.

    Legislation and rulemaking · European Union, AI-generated CSAM, European Parliament

  95. PIRG report finds AI toys claim to run on chatbots meant for adults and teens

    · NBC News

    A US PIRG Education Fund report found more than two dozen toys advertised as powered by OpenAI or Google models, despite age limits on those chatbots. PIRG got developer access from Google, OpenAI and xAI with no substantive vetting. Anthropic’s terms require users to be over 18.

    Research and evidence; AI chatbots and companions · AI toys, AI toys (devices), chatbots, Google, OpenAI

  96. UK to trial social media ban for children under 16

    · chinadaily.com.cn

    The United Kingdom is seeking feedback from parents and children through trials of a potential under-16 social media ban, along with curbs on gaming services and AI-powered chatbots.

    Age limits and social media bans; Lawsuits and settlements · social media ban, United Kingdom, chatbots

  97. MEPs ask Commission how it will respond to reports Meta smart glasses recorded intimate momentsAllegation

    · euractiv.com

    Members of the European Parliament submitted a written question to the Commission after Swedish newspapers reported on February 27 that Meta’s Ray-Ban AI glasses recorded users in intimate situations, with footage reviewed by a contractor in Kenya. Meta said content is filtered first. The allegations are unproven.

    Legislation and rulemaking · Meta, European Commission, European Parliament, European Union, Kenya

  98. Europe pressed to slow digital age-verification push amid privacy fears

    · euronews.com

    An open letter published March 2 from 371 security and privacy academics in 29 countries urged governments not to deploy online age checks until scientific consensus settles on their benefits and harms. They warned the checks risk exposing users to malware, scams and extra disclosure of personal data.

    Age assurance and verification; Privacy and children’s data · privacy and data protection, age assurance, settlement

  99. UK government proposed a power to bring AI chatbots under the Online Safety Act

    · GOV.UK

    In final amendments for the Crime and Policing Bill, the UK government proposed a power for the Secretary of State to bring unregulated AI services into the Online Safety Act’s illegal-content duties. It also tabled automatic data preservation notices for deaths of children aged five and over.

    Legislation and rulemaking; AI chatbots and companions · Online Safety Act (UK), United Kingdom, chatbots

  100. Singapore is studying whether to add safeguards for online games and AI chatbots

    · AsiaOne

    A Singapore minister told Parliament the government was studying whether safeguards on online video games and AI chatbots were needed. App stores must apply age assurance to block under-18 users from age-inappropriate apps from March 31.

    Legislation and rulemaking; AI chatbots and companions; Gaming and virtual worlds · chatbots, Singapore, Apple, age assurance, app stores

  101. Australia reviews children’s social media ban

    · iapp.org

    Australia has begun reviewing its ban on social media accounts for children under 16, introduced in December 2025. Australia’s eSafety Commissioner is tracking more than 4,000 children and families to assess how the policy works in practice.

    Age limits and social media bans; Legislation and rulemaking · Australia, social media ban, eSafety Commissioner

  102. Ministers consider social media curfew and AI chatbot restrictions to keep children safe online

    · independent.co.uk

    The UK government planned to open a consultation on March 2 on options including overnight curfews on children’s social media use, restricting children’s access to AI chatbots and a possible minimum age. It was set to close May 26, with a response planned for summer.

    Legislation and rulemaking; AI chatbots and companions · chatbots, Safe Online, age limits, United Kingdom

  103. UK launches consultation on possible social media ban for under-16s

    · bbc.com

    The UK government opened a public consultation on 1 March on whether to ban social media for under-16s. It also asks about switching off addictive features, overnight curfews, unrestricted chatbot use and stronger age verification. The consultation closes on 26 May.

    Age limits and social media bans; Legislation and rulemaking; Industry policy and partnerships · social media ban, United Kingdom, addictive design, age assurance, chatbots