AI Revolutionizes Racial Covenant Removal in Santa Clara County: A Step Towards Historical Justice

In a groundbreaking collaboration between Stanford University, Princeton University, and the County of Santa Clara, artificial intelligence (AI) has been employed to identify and remove racially restrictive covenants from millions of property deeds. This initiative marks a significant step towards addressing historical housing discrimination and ensuring compliance with California's anti-discrimination laws.

The problem of racial covenants, which prohibited individuals of certain racial backgrounds from purchasing homes, has persisted despite being unconstitutional since 1948. These clauses were often included in property deeds and remained embedded in historical records. The U.S. Supreme Court's ruling did not eradicate these discriminatory provisions, leaving them to continue affecting property ownership rights.

In 2021, California enacted a law requiring all 58 counties to proactively identify and redact racial covenants from property records. However, the sheer volume of documentsSanta Clara County alone has over 24 million deed recordsmade manual review impractical. The task was daunting, with Assistant County Clerk-Recorder Louis Chiaramonte noting that his team had to manually read nearly 100,000 pages over several weeks to find these covenants.

To address this challenge, Stanford University's Regulation, Evaluation, and Governance Lab (RegLab) partnered with Santa Clara County to develop an AI model. This collaboration utilized large language models to detect racial covenants with almost perfect accuracy. The AI tool was designed to search through documents from 1902 to 1980, a period when most of these racist covenants were created, and sift through 5.2 million deed records to identify racially restrictive language.

The results were impressive: the AI model identified over 7,500 deeds containing racial covenants, saving years of work and 86,500 person-hours. This not only expedited the process but also significantly reduced costs compared to hiring commercial vendors or relying on crowdsourced efforts. The County of Santa Clara has been actively reviewing millions of documents to eliminate discriminatory language from property records, with Chief Operating Officer Greta Hansen expressing gratitude for the partnership with Stanford, which has substantially accelerated this process.

The findings reveal a striking instance of racial covenant usage in Santa Clara County. By 1950, one in four properties in the county were subject to these covenants. The data indicates that the most affected groups were predominantly Black and East Asian populations, with 'Other Asian' groups ranking third. In contrast, Latino and European populations experienced the least discrimination.

The research also highlights that the majority of property deeds with restrictive covenants were labeled 'White-Only' properties, particularly prevalent in the 1920s. Additional covenants targeted other groups, including Italian, Portuguese, Indian, and Mexican individuals. This detailed analysis provides a comprehensive understanding of historical patterns of housing discrimination in Santa Clara County.

Stanford researchers have mapped 'White-Only' properties in Santa Clara County, illustrating clusters of racial covenants with significant developments highlighted in red and smaller subdivisions represented by dots. An interactive version of the map is available for further exploration.

The collaboration between Stanford University, Princeton University, and the County of Santa Clara demonstrates a path to using technology to aid in many other similar reform efforts. This initiative not only addresses the immediate need for racial covenant removal but also sets a precedent for leveraging AI in legal reform processes across jurisdictions.

Daniel Ho, a Stanford law professor and director of RegLab, led the project. Ho has a personal connection to the racially restrictive documents, having purchased a home in Palo Alto in 2015 where the historic deed included a clause prohibiting occupation by individuals of African, Japanese, Chinese, or Mongolian descent except as servants to a White person. This experience underscores his commitment to understanding local history and addressing ongoing issues related to housing discrimination.

The project's success has sparked interest in other jurisdictions. Washington, Minnesota, and Texas are among the states that have passed laws allowing owners to remove offensive or discriminatory language from historic home ownership records. However, these laws often place the burden of removal on homeowners, highlighting the need for more comprehensive solutions like the one developed by Stanford researchers.

Looking ahead, researchers are optimistic about collaborating with other jurisdictions pursuing similar initiatives. The model developed by Ho and his colleagues is being made available for free across the state and around the country to enable all jurisdictions faced with this task to identify, redact, and develop historical registers of racial covenants more effectively.

In conclusion, the use of AI in identifying and removing racial covenants in Santa Clara County represents a significant step towards historical justice. By leveraging technology to analyze decades of property records efficiently and accurately, this initiative not only addresses ongoing issues related to housing discrimination but also sets a precedent for future legal reform efforts.

Learn More at Direct Post