Guide-linked quiz · Medium Level

AI Chatbots, Children and Online Safety Laws Quiz

Tests COPPA parental consent rules, UK Online Safety Act age assurance duties, and California SB 243 companion chatbot protocols and lawsuits.

Question 1 of 8

Under COPPA, what must a covered service get before collecting personal information from a child under 13?

Preview every question8 questions

Answers and explanations appear as you take the quiz.

  1. Under COPPA, what must a covered service get before collecting personal information from a child under 13?
    • A signed statement from the child's school
    • Verifiable parental consent
    • A government ID from the child
    • Nothing, as long as the data is encrypted
  2. Which UK regulator enforces the Online Safety Act and has said generative AI chatbots can fall within its scope?
    • The Information Commissioner's Office
    • The Competition and Markets Authority
    • Ofcom
    • The Financial Conduct Authority
  3. Which requirement is part of California's SB 243 for companion chatbots?
    • Protocols to prevent content encouraging suicide or self-harm and to refer users to crisis services
    • A ban on all chatbots for anyone under 18
    • Licensing every chatbot through the state medical board
    • Storing all conversations on servers in California
  4. What does SB 243 require for users the operator knows are minors?
    • Showing ads only from approved vendors
    • Requiring a parent to sit in on every session
    • Deleting every chat after 24 hours
    • Periodic reminders to take a break and that the chatbot is not human, plus reasonable measures to prevent sexually explicit content
  5. Why is a "13+" checkbox often not enough for compliance?
    • Checkboxes are illegal in the UK
    • Many rules turn on actual knowledge, the likely audience, or reasonable measures, not just what users say about their age
    • COPPA applies only to users over 18
    • A checkbox counts as verifiable parental consent
  6. Why do self-harm safety systems score risk across a whole conversation, not only message by message?
    • Single messages are too short to be classified
    • The law forbids reading individual messages
    • Risk can build up gradually over many turns, especially in long role-play sessions
    • Scoring whole conversations is cheaper
  7. What was notable about the 2025 ruling in Garcia v. Character Technologies?
    • The court allowed most claims to proceed instead of dismissing the case
    • The case was dismissed because chatbots cannot be sued
    • The company was ordered to shut down
    • The FTC took over the lawsuit
  8. Which age assurance method does the guide describe as the weakest?
    • Age verification against an ID
    • Age estimation from face signals
    • Third-party age credentials
    • Self-declaration

Keep testing yourself

More quizzes picked for your level.

View all quizzes