Be Careful What You Google
The search engine creates a feedback loop that influences the way people see the world, a new study finds.
Google is a modern oracle, and a miraculous one at that. It can lead you to the Perfect Strangers theme song lyrics, or to a satellite image of your childhood neighborhood, or to a blueprint for building a quantum computer. But for as much as it is a portal to the world's knowledge, and despite its inherently aspirational functionality, Google searches are also a reflection of the status quo.
Do an image search for "CEO," for instance, and Google's algorithm returns a mosaic of mostly white, male faces.
Which makes sense: Only two dozen Fortune 500 companies have women as top bosses—that's less than 5 percent of overall Fortune 500 CEOs. And the 10 best paid CEOs in America are all white and male, according to The Guardian.
There is, however, one female face that pops up among the first few dozen male CEOS, though.
Can you spot it?
Let's zoom in a little.
That's right. It's CEO Barbie.
Here's the thing, though: Google image searches don't just reflect the sad state of diversity in corporate leadership; they actually influence the ways in which people think about what it means to be a CEO. That's according to researchers at the University of Washington and the University of Maryland, who determined that Google images measurably sway a person's opinion about how many men and women work in a particular field—compared with what that person thought before conducting the search. The effect is small but significant. "We find that
people’s existing perceptions of gender ratios in occupations are quite accurate," the researchers wrote. "But that manipulated search results can... [shift] estimations on average [by] 7 percent."
The researchers also examined Google accuracy to see if the search engine's visual representation of women in a given industry matched up with actual employment numbers. In many cases, they didn't. Female CEOs were significantly underrepresented in Google image searches compared with their real representation in the workforce, while female telemarketers were significantly overrepresented in Google image searches. From the study:
In a few jobs—including CEO—women were significantly underrepresented in Google image search results, the study found, and that can change searchers’ worldviews. Across all the professions, women were slightly underrepresented on average...
In some jobs, the discrepancies were pronounced, the study found. In a Google image search for CEO, 11 percent of the people depicted were women, compared with 27 percent of U.S. CEOs who are women. Twenty-five percent of people depicted in image search results for authors are women, compared with 56 percent of actual U.S. authors.
"I was actually surprised at how good the image-search results were," said Matt Kay, a computer scientist and the co-author of the study, in a release published by the University of Washington. "They might slightly underrepresent women and they might slightly exaggerate gender stereotypes, but it’s not going to be totally divorced from reality."
Except for what researchers call the "sexy construction worker problem." It turns out that even when women show up in search results often enough, representations of them are very often, well, ridiculous. "For example," researchers wrote, "we identified many examples of sexualized depictions of women who were almost certainly not engaged in the profession they portrayed; we dub this the sexy construction worker problem, as images of female construction workers in our results tended to be sexualized caricatures of construction workers."
Here's what you'll see if you image-search "construction worker":
Now look at what happened when I searched "female construction worker":
All this raises a question about what a search-engine algorithm ought to do. Should it challenge reality, or simply reflect it? People tend to think about the act of googling something as clinical, technological—decidedly not human. But search engines are designed by humans who have diverse value systems and distinct ways of categorizing their understanding of the world. (Not to mention all of the humans who are uploading and tagging images that Google's algorithm finds online.)
And those ideas help build algorithms that influence the way tons of information is presented to billions of people. Search engines aren't just the first place people turn for answers, but very often the only place. But Google, like the oracles that have come before it, can be more opaque than it appears.