Proxy discrimination occurs when a seemingly neutral variable (e.g., zip code) serves as a proxy for a protected characteristic, such as race, gender, or socioeconomic status. In this case, the AI uses zip codes to determine learning support for racial segments, introducing unintentional racial bias because zip codes often reflect historical segregation and economic disparities.

Since historical housing policies (e.g., redlining) led to racial and economic segregation, using zip codes as a decision factor unfairly correlates race with access to educational resources—a classic example of proxy discrimination.

Algorithmic Bias & Fairness (FAT/ML Guidelines) – AI systems must avoid using proxies that reinforce systemic discrimination.

Civil Rights Laws (Fair Housing Act, Title VI of the Civil Rights Act) – Policies that disproportionately harm racial groups violate anti-discrimination laws.

ACM & IEEE Ethics in AI – Developers must mitigate biases in AI models that result in unfair treatment.

Deontological Ethics & Fairness – AI systems should not perpetuate historical biases through indirect discrimination.

Relevant Ethical References in Technology:Thus, the correct answer is C. Proxy discrimination, as zip codes indirectly introduce racial bias into the AI system.