Agents in Video Games

End-of-lesson check

15 questions · take it digitally for instant feedback at tendril.neural-forge.io/learn/quiz/end-builders-agentic-agent-game-ai

1. Which sentence best captures the main idea of 'Agents in Video Games'?

   1. Agents should always run without limits or oversight
   2. Tools and goals are unnecessary for agent design
   3. Agents and chatbots are the same thing in every way
   4. Modern video game NPCs use AI to react more naturally — they remember conversations, change behavior over time, and feel more alive..

2. Which of the following is part of 'A typical AI-NPC behavior'?

   1. Hide tool calls from the operator
   2. Use the most expensive model regardless of fit
   3. Ignore cost when scaling
   4. An NPC remembers you helped them last week. This week they offer you a discount. Next week they are hostile because you sided with their rival.

3. Which of the following is part of 'Three things AI NPCs do'?

   1. Always run with no oversight
   2. Hide tool calls from the operator
   3. Use the most expensive model regardless of fit
   4. Remember player choices over time

4. Which of the following is part of 'Review date'?

   1. Reviewed in 2026. Treat fast-changing product names, prices, availability, and policy details as examples to verify before use.
   2. Never log what the agent did
   3. Ignore cost when scaling
   4. Always run with no oversight

5. What is 'NPC' in this context?

   1. A trick to bypass approvals
   2. A core concept covered in Agents in Video Games
   3. A reason to skip all logging
   4. A way to disable the agent's tools

6. What is 'memory' in this context?

   1. A reason to skip all logging
   2. A trick to bypass approvals
   3. A core concept covered in Agents in Video Games
   4. A way to disable the agent's tools

7. What is 'reactive AI' in this context?

   1. A core concept covered in Agents in Video Games
   2. A reason to skip all logging
   3. A way to disable the agent's tools
   4. A trick to bypass approvals

8. How is a modern agent in a game different from classic 'game AI' from older titles?

   1. Classic game AI used large language models
   2. There is no difference at all
   3. Modern agents can plan, use tools, and adapt to player behavior in more open ways
   4. Modern agents only follow scripted paths

9. Why is logging every tool call an agent makes a baseline requirement?

   1. Logs are needed to debug, audit, and explain agent behavior to users
   2. Logs are only for legal teams
   3. Logs replace the need for testing
   4. Logs make the model run faster

10. What is the safest first place to deploy a brand new agent?

    1. Inside a critical billing system
    2. Production, against real customers
    3. A sandbox or low-stakes task with reversible actions
    4. On a public server with no auth

11. Which is the clearest sign an 'agent' is really just a chatbot in disguise?

    1. It only produces text and never takes actions
    2. It uses a system prompt
    3. It can call a search tool
    4. It can remember last week's conversation

12. Before letting an agent take a destructive action, what is the safest default?

    1. Skip approvals if the user trusts the agent
    2. Require explicit human approval for the specific action
    3. Hide the action from any log
    4. Approve once and let the agent repeat forever

13. Why are clear success criteria critical when building an agent?

    1. Without them you cannot tell whether the agent worked or guess
    2. They make the agent sound smarter
    3. They are required by law
    4. They reduce the number of tokens used

14. What is the difference between an agent's memory and its context window?

    1. Context is what the model sees right now; memory persists across runs
    2. Memory is faster but less accurate than context
    3. Nothing — they are the same thing
    4. Context lasts forever; memory is cleared every minute

15. What is the most reliable way to keep an autonomous agent from going off the rails on a long task?

    1. Trust the model to know when to stop
    2. Set a clear goal, a step budget, and review checkpoints
    3. Run it for as many steps as possible without checking in
    4. Disable its tools so it can only think