Lines of Thought

Children, companions and deepfakes

How are governments responding to AI chatbots and image generators harming children?

Regulators moved from inquiries to binding duties in 2026, prompted by teen-harm lawsuits and the Grok image scandal. Approaches range from takedown duties and platform investigations to audits and outright bans on companion features for minors.

  1. Enforcement or rulingInformationUS
    FTC orders seven firms to explain how AI companion chatbots protect children

    The FTC demands information from seven chatbot providers on protecting minors.

  2. Rule · In forceUS-CA
    SB 243

    California's first companion-chatbot law requires AI disclosure and crisis protocols from 2026.

  3. Enforcement or rulingInformationUS
    Character.AI and Google agree to settle lawsuits over teen suicides and self-harm

    The first teen wrongful-death chatbot suits settle, keeping liability pressure on companion apps.

  4. Enforcement or rulingInformationGB EU
    Ofcom opens Online Safety Act probe into X over Grok sexualised deepfakes

    Grok's sexualised images trigger UK and EU investigations of X.

  5. Rule · Partly in forceGB
    Crime and Policing Act 2026 (nudification tools, AI CSAM generators)

    The UK criminalises supplying nudification tools and AI child-abuse image generators.

  6. Rule changeInformationCN
    China issues companion-AI rules banning virtual partners for minors

    China bars virtual-partner AI services for minors and mandates usage reminders.

  7. Enforcement or rulingInformationUS
    FTC begins enforcing TAKE IT DOWN Act's 48-hour removal duty for intimate deepfakes

    US platforms must remove intimate deepfakes within 48 hours of a request.

  8. Country
    Australia

    Australia's under-16 social media minimum age law (penalties up to A$49.5m) is the age-gating alternative to regulating features.

  9. Rule changeInformationUS-CA
    California enacts SB 1119 requiring child-safety risk assessments for companion chatbots

    California moves to pre-release child-safety risk assessments and audits for chatbots.

Where this line could go next

Connected developments not on this line