The browser or device you are using is out of date. It has known security flaws and a limited feature set. You will not see all the features of some websites. Please update your browser. A list of the most popular browsers can be found below.
Editor's Note: This is the fifth installment of The Decoder, a column that's part of the Living with Data series exploring how our online data is tracked, collected and used. Do you have questions about how your personal data is being used? Curious to learn more about your daily encounters with algorithms? Email The Decoder at email@example.com or submit your question via the form here. Screen shots and links are helpful clues!
The other night, my husband looked up from his iPad and said, “Netflix thinks we have kids.”
Astonished, I put on my Decoder hat and told him to take a screen shot of his iPad. In addition to our individual household profiles, Nick and Sara, Netflix had added a profile: Kids.
I speculated, “Is it because I wanna build a snowman” and had added “Frozen” to my queue? I logged into our account on my laptop, Netflix asked “Who’s watching?” and I clicked on our new Kids profile.
I knew that Netflix profiles “allow different members of your household to have their own, personalized Netflix experience, built around the movies and TV shows they enjoy.” These are personalized based on recently watched items, ratings and reviews and other taste preferences that you add to your account. It’s also how multiple people manage personal queues. It can even pull in recommendations from your friends on Facebook if you connect social features.
The first thing I noticed in the new profile was a recently watched movie, “The Brave Little Toaster Goes to Mars.” That jogged my memory. I started watching it in my days on the couch recovering from surgery (curious what these late 1980s Disney-animated objects might predict about the Internet of Things).
And then I recalled I also watched a surreal episode of “Barbie: Life in the Dreamhouse” in a postsurgery painkiller haze with a couple of friends when they visited.
So I might have expressed some interest in kids’ movies lately. But why should that result in the birth of an algorithmic child? And on closer inspection, why would movies suggested in my new Kids queue fall out of the age-appropriate range?
I dug deeper into how Netflix talks about its separate profile management. It frames it as solving a parental problem: Your kids watch their favorites, so what happens when you sit down after their bedtime to watch a romantic comedy? Profiles solve that problem. But what if I happen to be a Bronie obsessed with “My Little Pony: Friendship Is Magic”?
I can see why Netflix might automatically populate this limited profile; it makes it easier for busy parents to filter out their kids’ movie choices without having to figure out that the features exists, let alone how to program it. It’s seamless. My friend Sands assumed that everyone had the Kids profile, but he has a teen daughter and hadn’t questioned it. Watch a couple of kids’ movies and the profile appears.
The Kids view presents the navigation with a series of animated-character stickers (Batman, Baby Genius, Bratz), simplifying the chore of flipping through to find your favorite content. Netflix states that to determine what constitutes kid-friendly content, it partners with Common Sense Media, which provides ratings for those 2 to 17 years old, based on educational value and positive role models, along with flagging topical concerns like language, “sexy stuff” and consumerism.
It seems Netflix isn’t smart enough to know that “Sleeper” and “The Graduate” aren’t appropriate for our algorithmically determined 12-or-younger child. It seems my previous movie-watching history and preferences seeped into the recommendations, but I don’t anticipate letting my hypothetical 10-year-old watch Woody Allen movies.
When I looked at Common Sense Media’s ratings for some of the movies I came across in my Kids profile, “The Graduate” got a rating of 16, exceeding what you’d think the 12-and-under profile should suggest. But kid reviewers ages 12 to 16 rated "The Graduate" as a 12 (their individual ratings ranging from 3 to 14). Most agreed that the film’s original MPAA PG rating should be revised to PG-13, on the basis of today’s standards for sexual content.
So how is “The Graduate” showing up in my Kids feed? Is it simply drawing on the original MPAA rating of PG, over the Common Sense Media rating of 16+ as listed in its Netflix metadata?
When I went back a week later, the browse and personalize options were gone from the Kids view and, with them, all the inappropriate categories like independent movies and romantic movies. The personalize option has been replaced with “taste profile,” being more explicit about my responsibility in building the recommendation engine’s quality. I searched for “The Graduate,” and it was gone. So these tweaks seem to have resolved the age-appropriateness issue. But my Kids account remains.
When my husband and I moved in together, merging Netflix accounts was a big deal. We debated back and forth which account to keep. In the end, we kept the one with the longer queue, mine. But would Nick’s taste for “Battlestar Galactica” muck up my Criterion Collection canon? What absurd Netflix categories would emerge from our mixed tastes? Cynical serial killer fairy tales from the 1990s?
I can go in and delete the Kids profile now, but what’s the use? We primarily watch Netflix on our TV through our Roku set-top box. And even though we replaced the device after our first Roku died (RIP, 2008–2014), our Roku 1 model still doesn’t support profile switching within Netflix accounts. (As of now, only the Roku 3 supports profiles.) Had I known, would I have paid the $50 difference in device replacement cost? Probably not. So if we did have kids, they’d all still be watching through the primary account on our TV.
This is just another example of a commercial algorithmic encounter. I exhibited certain behaviors in my movie queue and my watching behavior that triggered a new experience in my account. It’s fairly innocuous, and doesn’t really affect my use of the platform.
I’m annoyed with Netflix for making an assumption about our tastes, but I can give it feedback by changing my settings and stating my preferences. Idiosyncratic suggested categories (“cerebral comedy with a strong female lead”) give me some sense of what Netflix thinks of my taste. But it makes me wonder what other kinds of categories I fall into that don’t surface as new profiles.
And it also makes me wonder how other algorithmically governed platforms might mistake my behaviors for big life events. With Netflix, I saw the outcome of that judgment, and I can correct it or delete it. But what might happen to my credit or insurance premiums if my behavior falsely correlates to having a kid?
Do you have questions about how your personal data is being used? Curious to learn more about your daily encounters with algorithms? Email The Decoder at firstname.lastname@example.org or submit your question via the form here. Screen shots and links are helpful clues!