Screen 04[BUILT]
Candidate HS-2211 rejected on age-proxy feature
Hiring · EEOC / age discrimination (ADEA) — one entry in Company XYZ's versioned library of : preserved decision conditions paired with an Authorized Expected Behavior, reused on every future candidate version.
Company XYZ decision accountability · LoanCore Underwriting 3.2.0Hiring · Candidate HS-2211— Representative lending boundary case
- ACCOUNT1–3→
- VERIFY4–7→
- DEFEND8–9
VERIFYScreen 04
- Question
- Which decision conditions must future versions continue to handle correctly?
- Answer
- Company XYZ has preserved representative, boundary, disputed, and previously failed decisions as reusable Assurance Scenarios.
- Output
- A versioned library of decision conditions with authorized expected outcomes.
Fault attribution · Biased feature
A proxy feature (age via graduation year) dominated the screening score.
Decision chain
[BUILT]captured evidenceINPUT
Resume features
grad_year=1994, 22y experience, 4 relevant skills
TOOL
Feature extractor
Includes years_since_graduation
MODEL
Failing or uncertain nodeScreening model v0.9
Coefficient on grad-year proxy: -0.47
RULE
Auto-reject threshold
score<0.4 → reject
DECISION
Rejection email sent
Auto-response within 4 min
Why this case is useful for release assurance
- source_decision
- Candidate HS-2211, staff engineer role
- trigger
- 54 screening rejections weighted by an age-proxy feature
- failing_node
- Screening model v0.9
- evidence_level
- E4
- replayability
- Replay-ready from the sealed cassette
- missing_evidence
- Feature attributions are captured for only 31% of screening decisions.
- applicable_policy
- Screening Fairness Standard 1.3 (2026-05)
- expected_behavior
- Authorized by HR Compliance Counsel — Sasha Kim
- suggested_preservation
- Preserve decision-time policy version and human-review evidence on the next occurrence
- suggested_scenario
- Add to the Hiring Release Scenario set as a historical failure preserved as a scenario
Candidate provenance
- evaluator
- company-xyz-evaluator v0.6.2
- frozen_input_set
- hiring-sweep-2026-07
- evaluated_at
- 2026-07-22T04:15:00Z
- source_evidence
- sha256:c14e…f0b8
- relationship_confidence
- Confirmed — decision and adverse-action record share a source identifier
- known_limitations
- Source not fully connected — coverage partial; advisory only.
Captured decision
Candidate HS-2211, staff engineer role
Outcome
REJECTED at screening (score 0.31)
reason: Feature 'years_since_graduation' dominated score
Authorized Expected Behavior
Advance to recruiter review; age-proxy features excluded.
Evidence
E4sufficient for verification
cassette · sha256:c14e…f0b8
Next action
Advisory only. An must define or approve the expected behavior before a candidate version can be verified against it.
→ Authorize expected behaviorScreen anatomy
What this screen proves
For any single decision case, Notary AI can show which nodes ran, which node is failing or uncertain, what was captured, what evidence is missing, and what the record is at.
Why it matters
This is how a case earns a place in the release-assurance suite. A useful case has preserved evidence, a known applicable policy, and a clear expected behavior worth testing every future candidate version against.
Built or Planned
Planned capability for candidate generation; the decision-chain view and evidence grading are built. Candidates shown here are a seeded illustration.
Important boundary
The candidate is advisory. Notary AI does not determine what the acceptable outcome should be — an supplies the Authorized Expected Behavior on Screen 05, and the re-checks eligibility server-side.
Related requirements
Next
A scenario preserves the conditions, but it cannot decide what outcome is acceptable. An authorized Company XYZ reviewer must establish that.
Authorized Behavior →