Responsible AI
AI may organize the evidence. A person owns the decision.
This product policy describes the intended boundary between assistance and human accountability. It is public so candidates, customers, and reviewers can challenge it.
Updated July 24, 2026Permitted assistance
Aptaryn may help extract, organize, compare, summarize, and draft information supplied through the product.
Assistance must retain source context, distinguish candidate-provided information from verified information, and expose meaningful uncertainty.
Prohibited decisions
Aptaryn must not independently reject, advance, rank as a final disposition, hire, fire, or otherwise make a consequential employment decision.
An employer remains responsible for the criteria it chooses, the evidence it reviews, and the decision it records.
Disclosure and correction
People should know when AI helped produce content or a recommendation. They should be able to inspect the relevant factors, correct inaccurate context, and request human review.
Limitations
Model output can be incomplete, outdated, or wrong. Missing evidence is not evidence of absence. Reviewers should ask for clarification before reaching a negative conclusion.
Questions and concerns
Responsible AI questions can be sent to hello@aptaryn.com. A production policy should add response times, escalation ownership, and jurisdiction-specific processes before general availability.