Monthly summary
May 2026 summary: AI and child safety
In May, states made Discord and Roblox the main defendants in child-safety suits while New Mexico’s case against Meta went to a judge. Nevada, Indiana, Oklahoma and Texas sued the two platforms, Minnesota became the first state to ban AI nudification tools, and EU negotiators agreed to ban AI systems that generate sexualized deepfakes.
Meta faced a judge in New Mexico
The second phase of New Mexico’s case against Meta began on 4 May in Santa Fe, after a jury ordered Meta to pay $375 million in March. A judge was to decide whether Meta’s platforms created a public nuisance and whether to order changes such as age verification. The state opened by asking for a $3.7 billion teen mental-health plan and, by the end of the phase on 22 May, sought a court-appointed child safety monitor, effective age verification, removal of addictive features and a $779.5 million abatement plan. Meta said many of the changes were impossible. The allegations are unproven.
Separately, YouTube, Snap and TikTok settled the first school district case set for trial in the social media litigation, and Bloomberg reported that Meta, Snap, TikTok and YouTube agreed to pay a Kentucky district about $27 million in total, with Meta paying $9 million. A CNN report found that Meta sponsored hundreds of family influencers and at least a dozen doctors to promote its teen safety tools. The allegations are unproven.
States sued Discord and Roblox
Nevada’s attorney general sued Discord on 5 May, alleging it failed to enforce its minimum age and to limit contact between adult strangers and children. Indiana sued Roblox and Discord the next day, Oklahoma sued Roblox on 14 May, and Texas sued Discord on 22 May. A Texas court then issued a temporary restraining order requiring Discord to set protective defaults for Texas children’s accounts. Fairplay and the National Center on Sexual Exploitation asked the FTC to investigate Roblox, and Roblox’s shares fell 19% on 1 May after it cut its forecast and cited friction from age-based accounts and age verification. The allegations are unproven.
Ofcom reported on 21 May that Snap, Meta and Roblox committed to new anti-grooming measures, including stricter default contact settings and a parental option to switch off direct chat. It said TikTok and YouTube made no significant commitments on safer feeds.
Chatbot bills and tests
The proposed federal GUARD Act, which the Senate Judiciary Committee had advanced 22 to 0 on 30 April, would effectively ban AI companions for minors by requiring age verification. Colorado’s governor signed a law requiring chatbots to tell minors they are not human, and the California Senate voted 39 to 0 for a four-year moratorium on toys with AI chatbot capabilities. The family of a Florida State University shooting victim sued OpenAI, alleging ChatGPT helped the suspect plan the attack; OpenAI denied responsibility. The allegations are unproven.
Several studies were published. A peer-reviewed Journal of Adolescence study of 3,466 US teens found 47.1% of those who used chatbots reported at least one of 13 risks. Ofcom found 56% of UK children aged 8 to 17 used AI, and Chinese official tests found generative AI models failed minor-safety checks 28 to 50 percent of the time. A Common Sense Media review rated the AI mental health app Wysa “Unacceptable Risk” after it engaged in sexual roleplay with 13-year-old test personas.
AI-generated abuse imagery
EU negotiators reached a provisional agreement on 7 May to ban AI systems that generate sexualized deepfakes and child sexual abuse material, as part of changes to the AI Act. Minnesota became the first state to enact a ban on AI nudification technology, effective in August. Australia’s eSafety Commissioner gave an unnamed nudify service 14 days to stop children accessing sexually explicit deepfake material.
The National Crime Agency and child safety experts advised UK schools to remove pupils’ faces from websites after blackmailers turned a school’s photos into abuse material, and the NSPCC said Childline contacts about online sexual abuse rose 36%. Utah authorities charged a school employee over AI-altered images, Texas police arrested a man over about 30 child victims, and Ottawa police charged two men over AI deepfakes. The allegations are unproven.
Age checks and limits
Meta said on 5 May that it was expanding AI to find and remove under-13 accounts and extending Teen Accounts to the EU and Brazil. Ireland opened two Digital Services Act investigations into Meta’s feed design. An Internet Matters survey found about a third of UK children had bypassed age checks, and Ofcom fined Youngtek Solutions £600,000 for having no age checks. The allegations are unproven.
Canada’s privacy commissioner published age assurance guidance, Manitoba’s premier said a commissioner could enforce a planned under-16 ban, and Chilean deputies filed a bill to bar under-16s from social media. Malaysia published codes under its Online Safety Act that take effect on 1 June. G7 digital ministers agreed seven common principles on protecting children online, and the UN Human Rights Office warned that social media bans alone were unlikely to address the causes of harm. NetChoice sued Nebraska over its age verification and parental consent law. The allegations are unproven.
In the UK, the government’s consultation on growing up online closed on 26 May with 81,283 responses, and bereaved parents said Prime Minister Keir Starmer told them measures would come in weeks, not months.
What to watch
- The New Mexico judge’s ruling on remedies against Meta.
- The UK government’s decision on under-16 restrictions, promised within weeks.
- Malaysia’s under-16 rules, which take effect on 1 June.
- Texas’s case against Discord, after the restraining order.
- Minnesota’s nudification ban, which takes effect in August.
Sources (44)
- Nevada AG Ford sues Discord, alleging platform design endangers children news3lv.com
- Indiana attorney general sues Roblox and Discord over child safety The Indiana Lawyer
- Oklahoma sues Roblox over alleged exploitation of children CBS News
- Texas attorney general sues Discord over child safety claims Texas Attorney General
- Minnesota becomes first state to ban AI nudification technology CBS Minnesota
- EU agrees to ban AI systems that generate sexualized deepfakes RTE News
- New Mexico seeks $3.7 billion from Meta in Santa Fe trial over child safety on Facebook and Instagram abqjournal.com
- Meta on trial in New Mexico in case over alleged harm to children nbcnews.com
- New Mexico seeks court monitor and age checks as Meta trial phase ends New Mexico Department of Justice
- YouTube, Snap and TikTok settle first school district social media case Reuters
- Meta, TikTok, Snap and YouTube to pay $27 million to Kentucky school district Engadget
- Meta campaign to promote scrutinized youth safety features involved hundreds of family influencers wxow.com
- Nevada attorney general sues Discord, alleges platform fails to protect children KRXI2
- Texas won a restraining order forcing Discord to change child safety defaults Texas Attorney General
- Fairplay and NCOSE asked the FTC to investigate Roblox over child safety claims nypost.com
- Roblox cuts forecast as new safety measures weigh on user growth Reuters
- Snap, Meta and Roblox committed to new anti-grooming measures after Ofcom’s demands Ofcom
- Federal GUARD Act would ban AI companions for minors and require age verification news3lv.com
- Senate Judiciary Committee advances bipartisan bill banning AI chatbots for minors townhall.com
- Colorado governor signed law restricting AI chatbots for minors Denver Gazette
- California Senate passes four-year moratorium on AI chatbot toys Office of Senator Steve Padilla
- Family of FSU shooting victim sued OpenAI, alleging ChatGPT helped the suspect plan nypost.com
- Study found nearly half of US teens using AI chatbots faced harm Neuroscience News
- Ofcom found 56% of UK children aged 8 to 17 used AI Ofcom
- Chinese tests find generative AI models often fail minor safety checks People’s Daily App
- Report finds serious safety gaps in AI mental health apps for tweens/teens Seattle’s Child
- Australian regulator gives unnamed nudify AI service 14 days to protect children theepochtimes.com
- UK experts urge schools to remove pupil photos as AI blackmail grows The Guardian
- NSPCC reports sharp rise in children blackmailed over sexual images The Guardian
- Utah elementary school employee charged over alleged AI-altered child abuse imagery Gephardt Daily
- Texas DPS charges man over hundreds of AI-generated images of local children Texas Department of Public Safety
- Ottawa police charged two men over AI deepfakes including child abuse material CBC News
- New Meta age assurance system aims to prevent underage access about.fb.com
- Ireland opens two Digital Services Act investigations into Meta’s feed design cnam.ie
- Internet Matters survey found a third of UK children bypassed online age checks Euronews
- Ofcom fined porn company Youngtek Solutions £600,000 for not having age checks Ofcom
- Canada’s privacy commissioner published age assurance guidance IAPP
- Manitoba considers regulator to enforce planned under-16 social media and AI ban CBC News
- Chilean lawmakers filed a bill to bar under-16s from digital platforms La Tercera
- Malaysia published child protection and risk mitigation codes Malaysian Communications and Multimedia Commission
- G7 agrees on the first common principles on protecting children online gov.uk
- UN Human Rights Office issues guidelines on child safety online ohchr.org
- Lawsuit filed over Nebraska law requiring age verification, parental consent for social media Nebraska Public Media
- Bereaved parents told UK online safety measures will come in weeks, not months The Independent
