When To Use Agents Ethically

End-of-lesson check

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

1. Which sentence best captures the main idea of 'When To Use Agents Ethically'?

   1. Agents are powerful — and ethical use depends on disclosure, consent, oversight, and bounded harm..
   2. Tools and goals are unnecessary for agent design
   3. Agents should always run without limits or oversight
   4. Agents and chatbots are the same thing in every way

2. Which of the following is part of 'Three ethical questions'?

   1. Run unbounded retries on any error
   2. Avoid taking any actions in the world
   3. Have I disclosed AI use to people affected? Is there a way to override AI decisions? Will I keep records of what the AI did and why?
   4. Use the most expensive model regardless of fit

3. Which of the following is part of 'Three places ethics matters most'?

   1. Decisions about other people (hiring, lending, parole)
   2. Approve all actions automatically
   3. Always run with no oversight
   4. Never log what the agent did

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

   1. Approve all actions automatically
   2. Run unbounded retries on any error
   3. Use the most expensive model regardless of fit
   4. Reviewed in 2026. Treat fast-changing product names, prices, availability, and policy details as examples to verify before use.

5. What is 'disclosure' in this context?

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

6. What is 'accountability' in this context?

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

7. What is 'audit' in this context?

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

8. Which ethical concern is most specific to autonomous agents, beyond chatbots?

   1. They can produce fluent text
   2. They can be expensive
   3. They take real-world actions whose consequences are hard to undo
   4. They can be biased

9. Which of these is the strongest indicator that an agent workflow is ready to scale?

   1. It runs without any logging
   2. It passes a repeatable eval, has cost in budget, and a rollback plan
   3. It worked once for one user
   4. It used the latest model

10. Which is the best way to think about an agent's 'autonomy level'?

    1. It depends only on the model size
    2. It is binary — fully autonomous or not
    3. It is a slider from suggestion only, to ask-then-act, to act-then-report, to fully autonomous
    4. Autonomy does not affect risk

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

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

12. Why does an AI agent need 'tools' such as a browser, calendar, or code runner?

    1. Tools make the model speak more naturally
    2. Tools shrink the context window
    3. Tools let the agent take actions in the world instead of only producing text
    4. Tools replace the need for any prompts

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

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

14. Why is it dangerous to give an agent access to your email and calendar without scoped permissions?

    1. Scoped permissions slow the model down
    2. Broad access means a single misstep can send the wrong message or wipe events
    3. It will refuse to work
    4. Scopes only matter for paid accounts

15. Which budget control most directly prevents runaway costs from an agent loop?

    1. A longer context window
    2. A hard cap on steps, tokens, or dollars per task
    3. A bigger model
    4. A friendly system prompt