A Stanford University project is shedding light on patterns of housing discrimination in the South Bay by harnessing the power of artificial intelligence.
Genevieve Singh works in the Santa Clara County recording office. Lawmakers in 2021 ordered every California county to expunge their property records to identify and redact racially restrictive covenants.
“No person of African, Asian or Mongolian descent shall occupy this property,” according to a property deed in the Santa Clara County Recorder’s Office archives. It’s that type of racist language that county officials need to identify.
“We would look at books or the digital reel to locate restrictive words,” Singh said.
It’s quite a job. The province has 24 million deed documents totaling 84 million pages, dating back to 1850.
“Once we find them, we will work with the county council to send that for approval. After it was approved, it was re-recorded,” Singh said.
But Singh was only one of two county employees who read through these documents. To comply with the law, some California counties have entered into agreements with commercial suppliers.
Los Angeles County, for example, hired an $8 million company to conduct the scan for seven years.
Stanford professor Daniel Ho began to wonder if there was a more efficient way to document these racial treaties. He bought a house in Palo Alto and was somewhat surprised to encounter the racially explicit language.
“It was obviously disturbing to read the language and realize this about the history of the house you’re going to live in,” Ho said.
Ho, the director of the university’s Regulation Evaluation Governance Lab, which aims to modernize government, began working on an AI language model to speed up the process.
Researchers trained the model using thousands of racial matches from across the country.
“AI really offers the opportunity to study this, map it and create a historical record,” says Ho.
The laboratory got to work and scanned 5.2 million records between 1902 and 1980. It took about a week. They estimate that in 1950, one in four homes in the county was subject to racial compacts.
“The presence of these racial covenants can have long-lasting consequences because of the kind of community they convey,” says Ho.
Researchers found that just ten developers were responsible for a third of the identified covenants, indicating that developers had a lot of control over how Santa Clara County was built.
The REGLab says the AI model will save the province more than 86,000 man-hours.
“It really sped up the process. Instead of us going through everything with a fine-tooth comb and looking for these words, they identified it, so now we can just do quality control and move on with that package to the provincial counsel to get the process completed,” Singh said.
Researchers also found that racial covenants excluded African Americans at the same rate as Asian Americans, even when African American residents were less than one-tenth the size of the Asian American population.
The team also found a striking example of a cemetery in San Jose that only recorded burial records for “white people.”
The REGlab says this contradicts the conventional historical narrative that racial covenants were only used between private parties after the Supreme Court banned state-based racial zoning.
The team has released the article online and is making the model available to enable all jurisdictions faced with this task to more effectively identify, edit and develop historical records of racial treaties.
Kenny Choi
