Skip to content
Contact
TRUST & LEGAL

Responsible AI

Practical expectations for responsible recruitment assistance—not a claim that automated outputs are unbiased.

Draft updated 24 September 2026 · Philippines
Review draft — not a compliance certification

This content describes the current product and proposed policies. It requires legal and operational review before being treated as a finalized production policy or contract. Operator: JobbhyAI, Philippines. Privacy, security, and service contact: hi@jobbhy.io.

People remain accountable

Use automation to organize work and surface evidence, not to remove accountability. Reviewers should have enough information, time, and authority to question or override a recommendation. A routine approval click without reviewing evidence is not meaningful oversight.

Job-related evidence

  • Use criteria demonstrably relevant to the role and review their weights.
  • Do not infer sensitive characteristics from names, photographs, accents, or unrelated personal information to rank suitability.
  • Do not treat missing résumé detail as proof that a candidate lacks a skill; seek clarification where appropriate.
  • Provide accessible processes and consider reasonable alternatives or accommodations where required.

Testing and monitoring

Before using a scoring or AI feature to influence selection, assess accuracy and potentially discriminatory outcomes in its intended context. Reassess after material changes to criteria, models, or applicant populations. Any fairness analysis involving sensitive attributes needs a lawful basis and separate access controls. No independent bias audit is claimed here.

Record and challenge decisions

Keep job criteria, relevant evidence, and the human decision distinguishable. Correct errors without erasing legitimate historical records. Escalate harmful or misleading output and pause the affected use when necessary. Candidates should be able to contact the hiring organization about concerns.

Limits of these principles

These principles do not substitute for an organization’s legal obligations, required audits, or a deployment-specific risk assessment. Deterministic rules and human decisions also need scrutiny; disabling an AI integration does not remove all hiring risks.