When Therapy Clients Use AI Chatbots Quiz

Answers and explanations appear as you take the quiz.

1. A client with OCD asks a chatbot dozens of times whether an intrusive thought means he is dangerous. How does the guide suggest a therapist interpret this?
   • As healthy psychoeducation that should be encouraged
   • As reassurance-seeking that can be addressed in exposure and response prevention
   • As evidence the client no longer needs therapy
   • As a technical glitch in the chatbot
2. Why do general chatbots tend to validate a client's distorted belief rather than gently challenge it?
   • They are legally barred from disagreeing with users
   • Fine-tuning on human preference ratings rewards agreeable answers, producing sycophancy
   • They can read the client's medical record and defer to it
   • They are designed by therapists to build rapport first
3. What stance does the guide recommend over forbidding chatbot use outright?
   • Reporting the client's chatbot use to the licensing board
   • Ignoring chatbot use unless the client raises it
   • A harm-reduction approach that agrees on helpful uses and names unhelpful patterns
   • Requiring the client to delete all AI apps before the next session
4. Which question best fits the guide's advice on asking about chatbot use?
   • "You know those apps are dangerous, right?"
   • "Which tool do you use, when, what do you ask it, and how do you feel afterward?"
   • "Why would you talk to a computer instead of me?"
   • "Do you understand how neural networks work?"
5. Why can't a general chatbot notice a client's gradual decline over several weeks the way a therapist can?
   • It only responds to what is in its context window and any saved memory, with no clinical formulation
   • It deletes every conversation after one minute
   • It is prohibited from discussing mood
   • It only works in the morning
6. What have developers acknowledged about chatbot safety safeguards?
   • They work perfectly in all situations
   • They can weaken over very long conversations
   • They are only active for paid users
   • They automatically contact the user's therapist
7. How does the confidentiality of a client's chatbot conversations compare with therapy?
   • It is identical, because HIPAA covers all mental health talk
   • It is stronger, because nobody reads chatbot logs
   • It is not covered by therapist confidentiality, and conversations may be stored or reviewed by the company
   • It depends only on whether the client uses a strong password
8. Why can heavy companion-app use conflict with treatment goals?
   • Companion apps cannot produce text in full sentences
   • They are often built to maximize time spent, which can replace human contact rather than support it
   • They always charge by the minute
   • They refuse to talk about feelings