What happens when you let a toddler use a social media filter? When you present her with a live feed of her face with dog ears, with a flower crown, with big eyes and full lips and a ski-jump nose so delicate even Barbie might wonder if it's really serviceable for breathing? Or when you let a 2-year-old see what he'll look like at 65, or as a planet, or as a cartoon skull, or as every disciple at the Last Supper?
I have young children, and so do many of my family and friends. We all have smartphones with apps, most often Instagram and Snapchat, with filters that allow users to augment reality in real time. And whatever our parenting philosophy about screen time and social media, we've discovered — as, I expect, have most parents of our generation — that kids are transfixed by filters. Toddlers and even infants will sit, riveted, to see their faces morph from one filter to the next. An intriguing filter can break a tantrum. A scary one can induce terror.
Should we be letting our children see this stuff? Can they find the line between filter and reality? I know the animated version of me dancing on my laptop keyboard is just a digital novelty. I know the large spider Snapchat sends crawling across my scalp doesn't exist. It's not always clear whether children know that too — click that spider link and you'll see a montage of kids apparently convinced there really is a tarantula on their heads. Some giggle. More scream or flee. One, off-screen, reaches an arm into view to smack the nonexistent spider off another child's face.
I took these questions to Dr. Jacqueline Woolley, a psychologist at the University of Texas at Austin whose research focuses on the conceptual development of young children, including their ability to distinguish between fantasy and reality. In a conversation over email, I asked Woolley what young children likely understand when they casually encounter augmented reality (AR) and if she thinks exposure to these filters poses a risk for child development.
There's a sense in which the filters are nothing new, Woolley told me. "It's essentially where Halloween costumes and face paint meet, with a high-tech gloss," she said. A bunny filter, like bunny face paint, usually won't make a child believe she has actually become a bunny.
But a marked emotional reaction to a filter may point to a child being "confused" or experiencing their "perception of reality [being] altered," Woolley continued. A child's emotions may be "highjacked by the stimuli and can evolve independent of [her] understanding of what's real." The kids in the spider video likely aren't exaggerating their panic, though many (especially the older ones) will understand, after their initial shock, that there's no spider outside the app. It's comparable, Woolley said, to an adult's experience of fear while watching a horror movie we know is fake.
For toddlers, however, the line between emotional response and comprehension of reality isn't so neatly drawn. "Research shows that children don't understand cases in which appearance conflicts with reality until 3 or 4," Woolley said. "If you ask a child about a sponge that looks like a rock, she will either say it's a sponge and looks like a sponge, or that it's a rock and looks like a rock. So a young child might not be able to state, if you asked him, that he really is a boy but looks like a bunny, but it's quite possible that he understands this on a more implicit level; in other words, it's highly unlikely that he would start hopping and eating carrots." The older the child and the less overwhelming the emotional reaction, the more certain we can be of that implicit awareness of what is and isn't real.
When researching this topic, I divided social media filters into two categories: AR filters, like the spider, and "pretty" filters, like the bevy of options to give you "Instagram Face." You know the look: "It's a young face, of course, with poreless skin and plump, high cheekbones," explained Jia Tolentino at The New Yorker last December. "It has catlike eyes and long, cartoonish lashes; it has a small, neat nose and full, lush lips. ... The face is distinctly white but ambiguously ethnic."
Instagram influencers create this face by hand — well, hand and FaceTune. They apply makeup, get plastic surgery, and digitally edit what cannot be "fixed" in real life. But for we in the masses who can't effect such a thorough transformation, some semblance of Instagram Face can be achieved via filter. I've got one such filter open on Snapchat right now. It's smoothed my skin, given it a slight golden glow I haven't earned from this summer's sun. My dark circles are gone, my nose and jawline different — more feminine? I can't pinpoint the change. My eyes are perfectly made up and topped with a cute pair of crayon-drawn glasses. The tweaks seem minor when the filter opens, but when it closes: revulsion. My brain needs a few minutes to catch up to accepting my real face again.
How do kids' brains handle that switch? Though bunny filters could go in either box, Woolley and I mostly talked AR. For insight into the pretty filters, I interviewed Dr. Christine Starr, a developmental psychologist who is now a postdoctoral scholar of education at the University of California, Irvine. Starr pointed to the gender dynamic of pretty filters: Young girls who once would have explored femininity by "copying their mother's makeup routine," Starr said, now might use a filter to see how their faces "look with a more 'grown up' feminine face, longer hair, or bigger lips."
"Exploration is normal," Starr continued, but the persuasive imagery filters provide can leave users, particularly girls, with a distorted perception of themselves. They may come to feel "negative or even dysphoric about their real, unfiltered bodies," especially if those bodies aren't close to the Instagram Face standard. In extreme cases, this dysphoria could end up in territory Tolentino's report examined, where plastic surgery is increasingly normalized as a way to bring filters to life.
My two categories, then, seem to have inverse risks: The AR filters will most effectively confuse the reality perception of very young children. By the time they start kindergarten, most kids will easily distinguish between their emotional response to a filter and their understanding that the filter isn't reality.
But the pretty filters grow more confusing with time and accumulated exposure. Filter-induced body dysmorphia isn't generated in a day, or, typically, in a toddler. We grow out of imagining ourselves with bunny ears; we don't grow out of wanting to be more attractive — even if the filter-inspired look we want is physically unattainable.
Both experts emphasized the importance of parents monitoring and managing their children's filter use. The "combination of parent provision of information and maturing cognitive abilities makes me confident that no lasting harm will occur from interacting with these AR filters," Woolley said. For my family, the limits we've already chosen for media use mean our kids won't encounter AR filters much, if at all, until they're old enough to understand and articulate what they're seeing.
Starr envisioned a longer and more challenging parenting project. She advised "age-appropriate conversations" centered on "open-ended questions that help their child think critically": What do you like about this filter? Is it funny? Does it make you feel pretty? For children old enough to have their own social media accounts: Do you feel uncomfortable posting an unfiltered photo of yourself?
"Conversations like this can be brief and should happen early and often," Starr said, and they should be informed by a child's individual habits: A "child using filters primarily to make themselves look older and 'sexy' or attractive is more concerning behavior than a child who is primarily using filters to take silly photos." And, as in almost any aspect of parenting, our own habits matter and will do at least as much as our conversations to set familial norms of phone and filter use. If I can only bear to share my filtered face, my children may conclude their faces need a filter, too.