Data/ML/AITable-stakes
Hallucination
Also written as AI Hallucination
When an AI model confidently generates false or made-up information that sounds plausible.
Think of it like
Like a confident student who, not knowing the real answer on a test, writes something plausible-sounding anyway rather than admitting 'I don't know.'
Junior or senior?
Junior sounds like
Treats hallucination as an unsolvable black box.
Senior sounds like
Can describe specific mitigation steps they took — grounding, citations, evals — and how they measured the improvement.
Ask them
“How did you measure or reduce hallucination in a system you shipped?”