# OPS-10 Quiz: AI Concepts

Take this after studying `notes/OPS-10-ai-concepts.md`.

Reply with answers like:
`1B 2A 3D 4C 5B`

## Questions

1. What is an AI hallucination?

A. A confident but false or unsupported output  
B. A verified source citation  
C. A backup rotation method  
D. A remote desktop protocol  

2. What should you do before entering private company data into an AI tool?

A. Post it publicly  
B. Disable all logs  
C. Check company policy and approval  
D. Ignore data sensitivity  

3. Why can AI output be biased?

A. Because backups are encrypted  
B. Because RDP is enabled  
C. Because toner is low  
D. Because output may reflect skewed training data or assumptions  

4. What should you do with AI-generated technical steps before using them?

A. Trust them automatically  
B. Verify them before acting  
C. Delete the ticket  
D. Skip testing  

5. Which AI use is highest risk?

A. Explaining a public error code  
B. Pasting sensitive customer data into an unapproved public AI tool  
C. Drafting a generic checklist  
D. Summarizing public documentation  

6. A technician uses AI to generate a troubleshooting script for a customer system. What should be done before running it?

A. Trust it and run immediately  
B. Test in a non-production environment and review each step  
C. Share it publicly for community feedback  
D. Skip testing since AI is accurate  

7. Which term describes AI that creates new content such as text, images, or code based on training data patterns?

A. Reactive AI  
B. Rules-based AI  
C. Generative AI  
D. Symbolic AI  

8. An AI tool recommends a cost-saving measure. Why should a human review it before acting?

A. AI recommendations are always wrong  
B. AI lacks judgment about organizational values, ethics, and context  
C. The recommendation will definitely violate the SLA  
D. The AI may have produced the wrong file format  

9. Which practice best protects against AI hallucinations causing a production issue?

A. Only use AI for creative writing  
B. Always verify AI-generated technical steps against official documentation before implementing  
C. Increase AI usage to reduce errors  
D. Disable logging when using AI tools  

10. What information should NOT be entered into a public AI tool?

A. General error messages  
B. Public documentation links  
C. Internal IP addresses, hostnames, credentials, or customer data  
D. The type of cable being used  

11. What is a prompt in the context of AI tools?

A. A command-line argument for a script  
B. The input or instruction given to an AI model to generate a response  
C. A type of firewall rule  
D. A backup rotation schedule  

12. A technician notices an AI tool gives different answers to the same question each time. What does this suggest?

A. The tool is malfunctioning and should be replaced  
B. AI models can be non-deterministic and outputs should be verified  
C. The technician needs to rephrase every question  
D. The AI has been compromised  

13. Which action aligns with responsible AI use in a corporate IT environment?

A. Using AI to process customer records without approval  
B. Following company AI policy, avoiding sensitive data, and verifying outputs  
C. Sharing AI outputs publicly to get feedback  
D. Bypassing IT approval to install an AI tool  

14. What does AI "training data" refer to?

A. The hardware specs of the AI server  
B. The dataset the model learned patterns from during development  
C. The user's prompt history  
D. The AI company's employee handbook  

15. A technician uses AI to draft a reply to a customer. What is the most important step before sending?

A. Send it immediately to save time  
B. Review the reply for accuracy, tone, and sensitive data before sending  
C. Ask the AI to resend it  
D. Delete the conversation history first  

## Answer Key For Instructor

1. A
2. C
3. D
4. B
5. B
6. B
7. C
8. B
9. B
10. C
11. B
12. B
13. B
14. B
15. B