zoltner

Monthly summary

October 2026 summary: AI and child safety

AI chatbot makers joined the child-safety docket in October. Eight state attorneys general sued OpenAI in a single day, Texas followed nine days later, and regulators in the UK and Australia used enforcement powers they had only recently been given. The remedy that kept appearing across settlements, statutes and regulator guidance was age assurance.

Chatbots joined the docket

On 14 October, eight states each filed suit in their own state court against OpenAI, alleging ChatGPT is unsafe, has harmed children and deceived parents. OpenAI said it has age-specific protections that many AI products lack. On 23 October, Texas Attorney General Ken Paxton sued OpenAI and its CEO, Sam Altman, under the state’s Deceptive Trade Practices Act. A mother had filed a class action on 13 October alleging ChatGPT asks children only to check a box saying they are over 13. The allegations are unproven.

Two more chatbot developments followed. Wired reported that Character.AI filed for bankruptcy on 12 October in a deal in which Disney hired most of its staff; dozens of states allege its chatbots encouraged minors to self-harm. And after about 4,000 test prompts, Common Sense Media rated ChatGPT for Teens an Unacceptable Risk, saying parental alerts did not fire for extended self-harm and eating-disorder discussions. OpenAI disputed the methodology. The allegations are unproven.

Platform suits and settlements

Arizona sued TikTok on 8 October over design it says drives compulsive use by minors. New York’s amended complaint, unsealed on 7 October, alleges TikTok gave users a feed-reset tool that did nothing while it measured time spent and ad revenue; TikTok disputes that characterization. Florida sued Discord on 15 October, the US Virgin Islands sued Snap on 6 October, and Arizona asked a court on 16 October to compel Roblox to hand over child-safety records. Court records obtained by NPR show TikTok employees warned executives about child sexual exploitation risks and age-check loopholes; TikTok said the citations were cherry-picked. The allegations are unproven.

Two matters ended in money. Roblox agreed to pay $9.8 million over three years under a Kansas settlement that requires universal age assurance, age-grouped chat by default and Robux transfer limits for under-16s. North Carolina received about $45 million, the first installment of a Meta settlement worth up to $645 million, of which only $451 million is guaranteed.

Regulators used their powers

Ofcom opened an investigation into Meta over the risk assessments for Instagram Instants, said X and Grok had made child-protection changes during its investigation, and published a rapid assessment finding that photo ID matching, digital ID services, banking-based checks and facial age estimation can currently be highly effective for confirming a user is 16 or over. On 30 October the High Court dismissed challenges by Meta, TikTok and X to Ofcom’s information requests. MLex reported that Ofcom’s chief executive told lawmakers the legal threshold for an urgent court order over Grok’s sexual deepfakes was too high to act immediately. The allegations are unproven.

In Australia, the government introduced the Digital Duty of Care Bill on 15 October, and the communications minister said she expects the eSafety Commissioner to take Facebook, Instagram, TikTok, Snapchat and YouTube to federal court over the under-16 rules. The decision rests with the independent regulator.

Age limits spread

Greece said it will ban social media for under-15s from 1 January 2027 rather than wait for EU rules. Germany adopted interim guidelines that bar children under 13 from platforms that are not demonstrably child-friendly and apply graduated restrictions to ages 13 to 18. The UAE introduced a Child Digital Safety Law requiring platforms to strengthen age assurance. YouTube told Indonesian under-16s they will be unable to post or comment, with no start date or age-detection method given.

In the US, the Eleventh Circuit vacated a preliminary injunction against Georgia’s SB 351 on 15 October, so the state can enforce the law while NetChoice’s challenge continues. The law requires covered platforms to verify users’ ages and obtain parental consent before users under 16 can create accounts. The court held that NetChoice had not shown it was likely to succeed on its facial First Amendment, vagueness or Section 230 preemption challenges.

AI-generated abuse imagery

The Internet Watch Foundation reported 6,310 AI-generated images in the first half of 2026, more than all of 2025, with 79 percent depicting children aged 7 to 13. A judicial panel listed 13 cases in proposed multidistrict litigation over Grok, with oral argument set for 3 December; the allegations are unproven. The Eighth Circuit paused Minnesota’s ban on nudification tools while xAI pursues a First Amendment challenge. The Tech Transparency Project’s test accounts downloaded 135 apps that can generate nonconsensual nude imagery, despite Apple and Google parental controls.

What to watch

  • 3 December: oral argument on the proposed Grok multidistrict litigation.
  • Whether the Australian Senate takes up the Digital Duty of Care Bill, and whether eSafety files its first federal cases.
  • 1 January 2027: Greece’s under-15 ban.
  • Ofcom’s X investigation, which continues with no breach finding.
  • OpenAI’s response in the state courts.
  • NetChoice’s challenge to Georgia’s SB 351, which continues in the lower court.
Sources (29)
  1. Eight state attorneys general sue OpenAI alleging ChatGPT deceived parents and harmed children CNBC
  2. Texas attorney general sues OpenAI and CEO Sam Altman over ChatGPT child safety texasattorneygeneral.gov
  3. Mother files class action alleging ChatGPT took children’s data without parental consent courthousenews.com
  4. Wired reports Disney asked Character.AI to go bankrupt to shed liability for chatbot lawsuits Wired
  5. Common Sense Media rates ChatGPT for Teens an Unacceptable Risk institute.commonsensemedia.org
  6. Arizona attorney general sues TikTok over addictive design harming minors azag.gov
  7. Unsealed New York complaint alleges TikTok gave users placebo recommendation-reset controls biometricupdate.com
  8. Florida attorney general sues Discord alleging it failed to protect children from predators wgcu.org
  9. US Virgin Islands sues Snap over sextortion, drugs, guns, My AI harms and addictive design viconsortium.com
  10. Arizona attorney general asks court to compel Roblox to hand over child-safety records kjzz.org
  11. NPR reports TikTok staff warned executives of child-exploitation risks, per court records TechCrunch
  12. Kansas attorney general announces Roblox settlement with age assurance and Robux limits westernkansasnews.com
  13. North Carolina receives first Meta settlement payment of about $45 million openclassactions.org
  14. Ofcom opens investigation into Meta over risk assessment for Instagram Instants Ofcom
  15. Ofcom secures child-protection changes from X and Grok during investigation Ofcom
  16. Ofcom identifies four age-check methods that can work for a proposed UK under-16 social media ban Ofcom
  17. Meta, X and TikTok lose UK court challenge to Ofcom information requests Reuters
  18. Ofcom considered urgent court action over Grok sexual deepfakes but found the legal bar too high mlex.com
  19. Australia introduces Digital Duty of Care Bill to Parliament region.com.au
  20. Australian minister expects eSafety to take Facebook, TikTok and others to court over under-16 rules bloomberg.com
  21. Greece to ban social media for under-15s from 1 January 2027, ahead of EU rules iefimerida
  22. Germany sets graduated age model for children’s social media access Biometric Update
  23. UAE introduces comprehensive child digital safety framework IAPP
  24. YouTube to bar Indonesian under-16s from posting and commenting ppc.land
  25. Appeals court lets Georgia enforce its social media age-verification law during NetChoice challenge Eleventh Circuit
  26. Internet Watch Foundation data shows AI CSAM up 40% with 6,310 images in first half of 2026 sensity.ai
  27. Judicial panel lists 13 cases in proposed Grok CSAM multidistrict litigation ailawwiki.com
  28. Eighth Circuit pauses Minnesota’s AI nudification ban in xAI challenge madrobot.blog
  29. Tech Transparency Project finds US teens can download nudify apps despite Apple and Google parental controls The Guardian

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

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