I crawled my ad settings to see what Google really knows about me

I apparently like a lot of things.

Google’s algorithms have come under fire once again. Researchers at Carnegie Mellon University and the International Computer Science Institute discovered recently that human-made computer algorithms are mildly judgmental.

To figure out just how discriminatory Google’s ad products can be, the researchers built AdFisher. The tool tracks the ads served by Google to the third-party sites across thousands of automated Web browsers it controls.

The researchers found that women see fewer job postings for high-paying jobs than men do after visiting job sites that make it clear to Google they’re in the market for a career change. Google’s ad settings, which are intended to let users edit the information it collects, neglects to mention some of the more sensitive information being collected. For example, Google displayed drug rehab programs to users who searched for substance abuse.

My ads settings page says I’m interested in 177 things based on my Google activity.

It’s important to remember that Google is just a vehicle for advertising. The ads targeting men more frequently than women could fall on the shoulders of advertisers, handpicking potential targets for job opportunities. 

But this isn’t the first time Google’s algorithms have displayed sexist or racist search results or descriptions. Earlier this year, Google was widely criticized when an image search for “CEO” showed mostly white men, and the first female result was Barbie. (A cursory search demonstrates this still hasn’t changed.) And last week, Google’s image recognition software categorized black people as “gorillas” in Google Photos.

As much as we might try to avoid it, Google is constantly sucking up our data in an effort to both enhance its own software, like image recognition and interest analysis, and to better serve advertising. Every bit of information we give Google—browsing history, search history, photos, and emails—is used to build a profile about us that follows us across the Web.

Thanks to the AdFisher study, I was reminded of Google’s ads settings page—the treasure trove of data that Google uses to show you ads across the Web. It’s a collection of topics that, based on your browsing history and other interactions on the Web, Google thinks you’re interested in. My ads settings page says I’m interested in 177 things based on my Google activity.

To be honest, I wish Google would send me more ads of beauty pageants; it would make advertising so much easier to ignore.

To figure out just how much Google actually knows about me, I took a granular peek at each of the 177 topics. Below is a breakdown of what my interests actually are, cherry picking some of the weirdest, broadest, and incorrect assumptions about me. 

Based on this information, Google definitely won’t be showing me ads for executive jobs anytime soon.

OK, Google

Google is supposed to be smarter than I am—as if my Web browsing history is a glimpse into my psyche that even I have not had the chance to dust off yet. And yet, some of these interests are so wrong and so random that I’m not entirely sure they should be considered something I want to actively spend time on, let alone click through an ad about them.

Anime & Manga
Unclear if this shows up because I watched My Neighbor Totoro or Googled tentacle porn.

Beverages
Why yes Google, I do enjoy ingesting liquid.

Bicycles & Accessories
I rode City Bike in New York for the first time recently, though I did not search for it first.

Dolls & Accessories
WTF, no.

Rabbits & Rodents
No idea where that came from.

Hunting & Shooting
Perhaps for the previous interest?

“Entertainment”

Of the 177 “interests,” a majority of them have to do with entertainment. “Arts & Entertainment” is a topic itself, and then Google provides more specificity around things it thinks I like.

There seems to be no rhyme or reason to it. It’s mostly things anyone who has a Netflix account or looks at things on the Internet would probably be into.  I don’t watch that much television; I avoid almost everything except HBO dramas and the occasional Netflix original series. And yet, these subjects are fair game for advertising to me.

TV Comedies
TV Commercials
TV Crime & Legal Shows
TV Dramas
TV Family-Oriented Shows
TV Medical Shows
TV Reality Shows
TV Sci-Fi & Fantasy Shows
TV Shows & Programs
TV Soap Operas

Computer games, board games, adventure games, and just “games” are all on the list as well, though I don’t particularly like games of any sort, board or otherwise. (I’m going to assume me being a tech reporter accounts for the computer and video bits of this list.)

Adventure Games
Board Games
Chess & Abstract Strategy Games
Computer & Video Games
Games
Roleplaying Games
Shooter Games

Spending money on things

I may or may not have a shopping addiction—and Google is aware of my habits. “Apparel,” is accurately listed as one of my interests, as the company is no doubt aware from the hundreds of emails with receipts from my favorite stores. “Shopping,” is understandably on there as well.

“Beauty,” “fitness,” and “beauty pageants” all appear, too. To be honest, I wish Google would send me more ads of beauty pageants; it would make advertising so much easier to ignore.

Accurate interests

Despite some interests that are completely off the reservation—small rodents, really?—Google does have a few things on this giant list of algorithmically generated interests that actually apply to my life.

Dance, air travel, dogs, books & literature, robotics, hobbies & leisure, and science are all topics I consume on the Internet on a regular basis. However, these, along with all the other hundred or so topics Google is trying to box me into, are quite broad and could apply to most average Internet users. People like being entertained, care about material items, and want to stay clean and healthy.

Perhaps Google can glean data like my sexual orientation, political affiliation, or religious beliefs from this dataset. But by analyzing it interest for interest, I appear to be an average, if not banal, person, with a wide variety of interests.

When your Web browsing history doesn’t actually matter

What it ultimately comes down to is judgement. And computers don’t have that yet.

AdFisher demonstrated that our data is shoved into a massive database, and it’s quickly analyzed by a piece of software that doesn’t know whether we actually like guns and hunting, or if we read a story about someone being fatally shot in a hunting accident. We could also scroll through endless amounts of hate-reading, and Google will take that as a signal that we’re interested in it. Algorithms, though programmed by humans, can’t make that distinction quite yet.

After taking a hard look at my ad settings page, I’ve realized there’s more obscurity in my Internet browsing history than I anticipated, and while Google’s massive collection of data makes me uncomfortable, this cursory look also makes me think that perhaps it doesn’t know me on a very human level.

To Google, I’m not a female human with actual interests. I’m just another data set that clicks on links. 

Photo via Robert Scoble/Flickr (CC BY 2.0) | Remix by Max Fleishman 

Selena Larson

Selena Larson

Selena Larson is a technology reporter based in San Francisco who writes about the intersection of technology and culture. Her work explores new technologies and the way they impact industries, human behavior, and security and privacy. Since leaving the Daily Dot, she's reported for CNN Money and done technical writing for cybersecurity firm Dragos.