Testing generative AI · Question 55 of 60
A healthcare chatbot is to be checked for safety using red teaming. What does this aim at, unlike security red teaming?
1 answer is correct. Hard
Show answer
Answer
- B How the system produces harmful outputs such as unsafe advice under ordinary use without adversarial intent
Explanation
Security red teaming is about malicious inputs. With safety and other goals, on the other hand, you often ask whether the system can cause harm during completely normal use, for example with a dangerous health tip, even though the user meant no harm.