Stanford Lab Uses AI to Remove Racially Restrictive Language from Santa Clara Property Deeds

A team of scholars led by Stanford University’s Regulation, Evaluation, and Governance Lab (RegLab) have partnered with the County of Santa Clara, California to remove racist language in millions of property deeds. The state of California enacted a law in 2021 that requires all counties to identify and redact any existing deed records that contain racially restrictive language. According to the Stanford project, the County of Santa Clara has over 24 million deed documents dating back to 1850 that contain racist verbiage that restricts individuals of African and/or Asian descent from owning property. To address the substantial volume of needed redactions, scholars at Stanford’s RegLab trained AI language models to identify deeds with racial covenants and developed a method of cross-referencing historical maps to locate relevant properties. By utilizing these technologies, the team at Stanford estimates their methods will save thousands of hours of human labor at a significantly reduced cost. In their analysis, the research team found that approximately one in four properties in Santa Clara were subject to racial covenants. The Stanford RegLab’s findings and an interactive map can be accessed here. The post Stanford Lab Uses AI to Remove Racially Restrictive Language from Santa Clara Property Deeds...