Responsible AI and Intelligence Policy
Our principles for human oversight, validation, privacy, and lawful AI and intelligence use.
Core principles
PQL prioritizes lawful and beneficial technology uses, clearly scoped purposes, evidence quality, safety, explainability, privacy, and human control. Consequential decisions should not rely solely on AI output without appropriate assessment.
Assessments, interpretation, and fairness
Models used in individual assessments, including NOEMA-related approaches, require fit-for-purpose scientific support, reliability and validity evaluation, bias and fairness review, appropriately limited claims, and ongoing testing. Clinical conclusions or employment decisions should not be inferred from a single insufficiently supported indicator.
Human control and approvals
For security response, information gathering, analysis, or high-impact actions, PQL emphasizes system-owner authorization, least privilege, decision records, and human escalation before risky actions.
Data and confidentiality
Client data, assessment outputs, confidential material, and investigative records must be handled within agreed purposes, permissions, access controls, and retention rules. Sensitive materials must not be sent to external services without authorization and appropriate risk assessment.
Technology limits and claims
PQL does not present AI models as invariably accurate, bias-free, or risk-free. Claims of maturity, certification, and performance should be made only with appropriate evidence, testing, and publication approval.
Use our official contact for inquiries, personal data rights requests, or security reports.
info@pancaindera.web.id ↗