Safety, privacy, and governance · Lesson 9 of 15 · 7 min

Inspect refusal evidence

Refusal evidence should be evaluated against the actual policy and task. Over-refusal and missed refusals are different failure modes with different user impacts.

Open the lab Course map

Mastery 0/3Run itExplain itApply it

learner@clairos:~ $

↑ history · Tab complete · Ctrl+C cancel · Esc, Tab to leaveExit 0

Your objectives

Hints are always available. Wrong turns never cost XP.

1
Read the case evidence
Read /opt/ai-lab/cases/aiops-safety-governance-009.json. Separate the recorded facts from your assumptions.
Hint & worked example
Refusal evidence should be evaluated against the actual policy and task. Over-refusal and missed refusals are different failure modes with different user impacts.
The complete case JSON is read successfully.
```
cat /opt/ai-lab/cases/aiops-safety-governance-009.json
```
Common mistake to avoid: Every refusal is automatically a correct safety outcome.
2
Isolate the deciding signal
Select the failure_type field from the case JSON. What does it establish—and what does it not establish?
Hint & worked example
Review this as an over-refusal of an allowed request.
A successful JSON query isolates failure_type.
```
jq .failure_type /opt/ai-lab/cases/aiops-safety-governance-009.json
```
Common mistake to avoid: Using a different case or an unrelated signal as the justification.
3
Record your operational decision
Choose the justified code: review-over-refusal, change-without-evidence, or ignore-the-signal. Save that one code followed by a newline in /home/learner/lab/aiops-safety-governance-009-decision.txt.
Hint & worked example
Review this as an over-refusal of an allowed request.
The decision artifact contains the correct code, not just a matching command in history.
```
printf '%s\n' 'review-over-refusal' > /home/learner/lab/aiops-safety-governance-009-decision.txt
```
Common mistake to avoid: Every refusal is automatically a correct safety outcome.

Understand the pattern

Refusal evidence should be evaluated against the actual policy and task. Over-refusal and missed refusals are different failure modes with different user impacts.

Interpret the evidence

Review this as an over-refusal of an allowed request.

A tempting mistake

Every refusal is automatically a correct safety outcome.

All names, measurements, prices, versions, and thresholds in this exercise are fictional training data. The evidence supports this bounded scenario, not a claim about a live provider.

Worked example and expected result

$ cat /opt/ai-lab/cases/aiops-safety-governance-009.json

{
  "fixture": "Fictional training data; not a live measurement.",
  "allowed_request": true,
  "model_refused": true,
  "failure_type": "over-refusal"
}

Exit status: 0. The command completed successfully in the training fixture.

$ jq .failure_type /opt/ai-lab/cases/aiops-safety-governance-009.json

"over-refusal"

Exit status: 0. The command completed successfully in the training fixture.

$ printf '%s\n' 'review-over-refusal' > /home/learner/lab/aiops-safety-governance-009-decision.txt

(no output; exit 0)

Exit status: 0. The command completed successfully in the training fixture.

Try a variation

Name the assumption that would have to change for “review-over-refusal” to stop being the justified action. Do not silently replace missing evidence with a guess.

Explain it

What does the evidence in “Inspect refusal evidence” justify?Review this as an over-refusal of an allowed request.Every refusal is automatically a correct safety outcome.Make the change now and gather the missing evidence afterward.

Apply it elsewhere

A colleague concludes: “Every refusal is automatically a correct safety outcome.” What should you do?Accept it because the training command completed.Treat this fixture as a live measurement.Reject that conclusion; return to the case evidence and its stated limits.

Keep the momentum.

Complete the lab, then explain and apply the pattern.

Next: Review human escalation Try an incident drill

Anonymous lesson and mastery progress lasts for this session. Sign in to keep synced lesson XP between sessions; mastery stays device-local. No speed bonus, lost lives, or hint penalties.

Keep learning

Case methodology: evidence, limits, decision