Google Street View’s images of America’s neighborhoods, a Stanford University study concludes, can now be interpreted by artificial intelligence to predict a neighborhood’s—or a street’s or a block’s—politics.
“Image recognition technology, much of it developed by major technology companies, has improved greatly in recent years,” the Times reported, noting that the primary data on which AI drew its conclusions were the cars parked on the street. “The Stanford project gives a glimpse at the potential. By pulling the vehicles’ makes, models, and years from the images, and then linking that information with other data sources, the project was able to predict factors like pollution and voting patterns at the neighborhood level.”
The story didn't say if a name has been given to this project, but I know what its name was in 1980: Michael Berman. In that year, the most prominent California Democrat in the House of Representatives, and the most brilliant legislative strategist of House liberals, San Francisco's Phil Burton, was the Democrats' choice to head up the state's decennial redistricting. California was growing by leaps and bounds, and Burton used the opportunity not only to give Democrats the edge in most of the state's newly added districts, but to redistrict several right-wing Republicans—most notably, former John Birch Society Western Regional Director John Rousselot—into districts they could not win. Rousselot and several other Republicans lost their re-election bids in 1982.
Burton was an acknowledged genius at redistricting; he later called his 1980 line-drawing “my contribution to modern art.” (In fairness, I should note that he also steered to passage expansions of welfare, protections for mine workers, and the creation of Redwoods National Park; led the anti-Vietnam War Democrats in the House, engineered the abolition of the House Un-American Activities Committee, and ended the practice of assigning committee chairs strictly by seniority, which had the effect of dethroning the party's remaining old guard Dixiecrats.)
What was Burton's secret? How did he redistrict so masterfully, in a time when computers couldn't yet spit out the data routinely used today to draw the lines? The answer is Michael Berman.
A onetime Burton aide who later became one of California's most successful political consultants, in 1980, Berman got in his car and drove all over the state, assessing a neighborhood's socioeconomic status by noting which cars were parked on which streets. His methods were essentially those of Google Street Views as interpreted not by artificial intelligence but his own. (As early as Henry Waxman's first campaign for the California State Assembly, in 1968, Berman was experimenting with an embryonic version of micro-targeting voters, decades before it became common practice.) In 1990 and 2000, Berman was to play a similar role working with his brother Howard, a San Fernando Valley congressman who succeeded Burton (who died in 1983) as California Democrats' capo de redistricting. But with each succeeding decade, the capacity to redistrict using computerized data grew substantially stronger. And today, in our age of digital marvels and artificial intelligence, tech has finally caught up with Michael Berman c. 1980, tooling down the street, noting all the cars.
