Gids-gekoppelde quiz · Middelmatig Niveau
Wanneer therapiecliënten AI-chatbots gebruiken Quiz
Er wordt getest hoe cliënten kunnen worden gevraagd naar het gebruik van chatbots, waarom sycofantie en geruststellingslussen riskant zijn, en welke houding artsen kunnen innemen om de schade te beperken.
Vraag 1 van 8
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?
Bekijk elke vraag8 vragen
Antwoorden en uitleg verschijnen terwijl je de quiz doet.
- 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
- 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
- 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
- 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?"
- 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
- 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
- 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
- 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