According to DeepLearning.AI, researchers from Stanford and Princeton have fine-tuned a language model to identify racially discriminatory clauses hidden within millions of historical property deeds in Santa Clara County, California. By leveraging advanced AI natural language processing techniques, the project automates the detection of restrictive covenants that prevented people of color from owning or occupying specific homes. This practical application demonstrates AI's growing role in historical document analysis, offering scalable solutions for legal compliance and social justice initiatives. The approach not only improves efficiency over manual reviews but also opens new business opportunities for AI-powered regulatory and real estate services. (Source: DeepLearning.AI, June 30, 2025)