Quiz: Can I Trust What It Says? 3 questions · 80% to pass 1. How does the lesson characterize hallucination?A temporary bug the next model release will eliminateA rare event that only affects free consumer toolsA user error caused by poorly written promptsA permanent property of the technology that working systems assume and gate forNewer models fabricate less, but no release takes the rate to zero. Trustworthy systems assume error and build verification into the process.2. Lawyers were sanctioned for filing briefs with fabricated AI-generated citations. What does the lesson identify as the actual process failure?They used a consumer tool instead of an enterprise oneThe AI output reached a court without any human verification gateThe model they used was outdatedThey failed to instruct the AI to be truthfulThe model behaved exactly as models behave. The damage came from a process with no gate between the draft and the court.3. According to the lesson, what is the appropriate verification gate for data work?Trusting the batch if the first record looks rightAsking the AI to double-check its own outputSpot-checking a sample of records against the source systemReviewing only the records a customer complains aboutSampling against the source catches systematic errors without re-doing the whole job. Asking the AI to check itself just produces more confident output. Check answers Retake quiz Back to lesson Next lesson →