Stanford University professor Daniel Ho, an expert on artificial intelligence and regulatory policy, detailed work his organization had done in California, where local counties were grappling with a state law that required each of the county recorders to review all of their historical deed records and identify racially restrictive covenants.

“And some counties approached this just by throwing human beings at it, Ho said. The problem is, in Santa Clara County, that’s 84 million pages of records.”

Ho said a simple AI system was developed that could quickly identify all the provisions for the restricted deeds, significantly reducing the amount of labor required to complete the task.

But looking toward the future, Ho said his hopes would be that there is an essential framework developed for governments “to be able to distinguish AI snake oil from AI to solve real public problems.”