Testing AI Bias in Southeast Asian Hiring

A regional investigation into how large language models (LLMs) respond to identity cues in Southeast Asian hiring scenarios. Through 50 controlled comparisons, the study reveals how changes in race, gender, education, geography, and company prestige influence model reasoning, even when qualifications remain constant.

The framework introduces a structured approach to evaluating AI fairness, emphasizing shifts in tone, logic, and confidence rather than outcome alone. Designed for high-context environments, the material offers a replicable rubric, test structure, and scoring system to support bias detection in real-world applications. Grounded in local nuance, this analysis informs safer, more accountable AI deployment across the region.