Responsible AI principles · Question 10 of 12
A chatbot for medication questions must not recommend dangerous doses. Which measures serve reliability and safety?
2 answers are correct. Medium Created with AI
Show answer
Answer
- A Monitoring responses in operation and analyzing anomalies
- C Testing with targeted edge cases and adversarial prompts before launch
Explanation
Reliability and safety require testing before release, including deliberately difficult input (red teaming), and ongoing monitoring afterwards. Translations promote inclusiveness, and deleting histories supports privacy.