Question 1
Which of the following best describes the concept of specification gaming in AI safety?
Question 2
In the context of AI safety, what is the primary concern regarding unintended side effects?
Question 3
What is the main challenge in ensuring transparency in complex deep learning models for AI safety?
Question 4
Which of the following best describes the concept of reward hacking in reinforcement learning, as it relates to AI safety?
Question 5
When an AI system is designed to operate within acceptable bounds, what is a primary consideration?
