This document reflects ATLAS's current governance model — terms may evolve as the programme matures.
Data Governance & Responsible AI
Our commitment to building a transparent, unbiased, and secure data ecosystem for the Southeast Asian AI landscape.
Core Principles
Regional Representation
Ensuring linguistic data reflects the true cultural nuances of Southeast Asia, not just direct translations.
Privacy by Design
Automated and manual PII (Personally Identifiable Information) scrubbing before any data enters our public repository.
Bias Mitigation
Multi-layered validation steps specifically designed to identify and remove harmful societal biases.
Anonymisation Protocol
Every contribution undergoes a strict three-stage PII scrubbing process before it is integrated into the Project SEALD datasets.
- Rule-based pattern matching (Regex) for IDs & Numbers
- Named Entity Recognition (NER) for sensitive locations
- Community Validation double-blind review