แบบทดสอบที่เชื่อมโยงกับคำแนะนำ · ปานกลาง ระดับ
คำจำกัดความของพระราชบัญญัติ AI ของสหภาพยุโรปสำหรับแบบทดสอบระบบ AI
ตรวจสอบความเข้าใจของคุณเกี่ยวกับคำจำกัดความของ eu ai act ของระบบ ai ขอบเขต ความรับผิดชอบหลัก และข้อจำกัดในทางปฏิบัติในกรอบการทำงานของ EU
คำถาม 1 ของ 8
Which feature is central to Article 3(1)’s definition?
ดูตัวอย่างทุกคำถาม8 คำถาม
คำตอบและคำอธิบายจะปรากฏขึ้นเมื่อคุณทำแบบทดสอบ
- Which feature is central to Article 3(1)’s definition?
- The system infers how to generate outputs from inputs
- The software has a subscription price
- The vendor labels it “smart”
- The interface contains a chatbot
- Which output is explicitly within the definition’s examples?
- A printed instruction manual
- A product warranty
- A recommendation
- A static company logo
- Does the Act require every AI system to adapt after deployment?
- Yes, all systems must continually retrain
- No; it says the system may exhibit adaptiveness
- Yes, only chatbots count
- No, autonomy is excluded from the definition
- What does varying autonomy mean in the definition?
- Every AI system must make decisions without people
- Only robots qualify
- The system must be fully autonomous
- Systems can operate with different degrees of independence from human involvement
- Which is the clearest contrast to the AI-system definition in the recitals?
- A model that classifies examples
- A recommender that infers rankings
- Software based solely on human-written rules that automatically executes operations
- A system that generates predictions from input
- Does an AI Act definition alone tell you that a system is high-risk?
- No; risk classification requires additional scope and use-case analysis
- Yes, every AI system is high-risk
- Yes, any generated text is high-risk
- No, the Act has no risk categories
- Can a product contain both AI and conventional software components?
- No, a product must be entirely AI or entirely non-AI
- Only if the product is open source
- No, every calculation is AI
- Yes; analyze the relevant component and system boundary
- Which facts are useful when classifying a borderline feature?
- Rely on the marketing name
- Document its inputs, outputs, objective, autonomy, and inference behavior
- Count the lines of code only
- Assume every automated action is AI