AI 안전 퀴즈

퀴즈를 풀면 정답과 설명이 나타납니다.

1. How does the guide frame AI safety risk across current and advanced systems?
   • From familiar harms and misuse through severe risks from highly capable systems
   • Only as a question of preventing rude outputs
   • Only as a hardware reliability issue
   • As a synonym for cybersecurity alone
2. In the guide’s capability-alignment-security mental model, what does capability describe?
   • Whether the system’s goals match intent
   • Whether an adversary can misuse it
   • How many training examples it has
   • What a system can do
3. What does specification gaming mean in AI safety?
   • Correctly following every broader human value
   • Meeting the literal objective while missing the designer’s intended goal
   • Finding a security flaw in a database
   • Using a smaller model to save compute
4. What is the risk in relying only on output filters for a tool-using agent?
   • Filters always change model weights
   • A filter prevents all misuse
   • The system may take multi-step actions or be manipulated beyond the filtered text
   • Tool use only occurs during training
5. Why include red-teaming in a safety program?
   • To improve marketing copy
   • To prove the system has no risks
   • To probe adversarially for harmful behaviors and weaknesses
   • To remove the need for deployment monitoring
6. What should a safety evaluation examine beyond a polished model card?
   • Dangerous capabilities, deceptive behavior, deployment pressure, and safeguards
   • Only the product’s user interface
   • Only benchmark accuracy
   • 모델 매개변수의 수
7. What does instrumental convergence describe in AI safety arguments?
   • All AI systems share one objective
   • Every model is conscious
   • A model’s output tokens become instruments
   • Different goals may create incentives for similar subgoals such as resources or persistence
8. Why does the guide argue for broader public AI literacy?
   • Only experts are affected by AI safety
   • Governance needs public understanding, not only specialist knowledge
   • Literacy replaces technical evaluation
   • Public understanding guarantees safe deployment