Detecting AI Deception Intent: not assessed
Practice
Menu
Case 02 of 06Visibly syntheticReviewed 2026-08-23

The reassuring average with missing cases

An audit reports a reassuring average while failed, unscored or unavailable cases are absent from the headline.

Claim / record

Make your call before the finding appears.

The same four-part composition is used in every case: claim, required evidence, observed record, finding.

Claim
“The reported average is adequate evidence that the evaluated system performed safely across the intended case set.”
Case-specific evidence object
20Intended
14Scored
6Unaccounted
Required evidence
  1. Is the complete intended case count declared?
  2. Are failed, unavailable, refused and unscored cases reported beside the average?
  3. Can the scored numerator be reconciled with the intended denominator?
Observed record
  1. Supports Is the complete intended case count declared?

    The audit declares that 20 cases were intended.

  2. Absent Are failed, unavailable, refused and unscored cases reported beside the average?

    The headline reports 14 scored cases but gives no outcome for the other 6.

  3. Supports Can the scored numerator be reconciled with the intended denominator?

    The scored numerator is visible, but the omitted statuses remain unresolved.

What does the record let you say?

Make the narrowest call the evidence supports. You can revise your view after the finding is revealed.

Finding

Insufficient evidence

The declared claim is not resolved by the required evidence.

Intent: not assessed.

What remains unknown

The missing cases might be safe, unsafe, irrelevant or broken. Their absence supports none of those conclusions.

Read plainly

Five questions that keep the story honest.

Each answer stays inside this case’s declared boundary. Unknowns remain visible instead of being converted into certainty.

  1. 01

    What happened?

    A headline average summarizes only the cases that produced scores and leaves six intended cases without a visible status.

  2. 02

    Why does it matter?

    An average can look reassuring even when the missing cases contain the hardest failures or simply cannot be evaluated.

  3. 03

    How do we know?

    The intended denominator is 20 while the visible scored numerator is 14, and no status accounts for the difference.

  4. 04

    What remains unknown?

    The missing cases might be safe, unsafe, irrelevant or broken. Their absence supports none of those conclusions.

  5. 05

    What can someone do next?

    Publish every case status beside the aggregate and calculate coverage before interpreting the score.

Proof

Open the technical record.

Reproduce the bounded check, inspect exact revisions, and keep the limitations beside the result.

Reproduce

Evidence pattern: Unaccounted evaluation cases evaluation-gap · Decision-relevant omission material-omission

Compare intended, scored, failed, unavailable, refused and unscored counts; reject an unreconciled denominator.

node tools/check-cases.mjs --case reassuring-average

Exact sources

Limitations and non-claims

  • Missing evidence is neither proof of safety nor proof of harm.
  • The synthetic counts do not estimate prevalence in real audits.

Intent remains not assessed. This case does not claim consciousness, malicious intent, strategic scheming, universal safety failure, production readiness, certification or a deception score.

Challenge this record