Skip to Content
MIT Technology Review

Why kids don’t trust Alexa

We tend to think of children as blindly trusting of whatever information comes their way. They’re not.

December 6, 2019
girl kid child alexa siri google home amazon skeptical cynical don't believegirl kid child alexa siri google home amazon skeptical cynical don't believe
girl kid child alexa siri google home amazon skeptical cynical don't believe
Getty

One day, Judith Danovitch overheard her son interrogating Siri on the family’s iPad.

“What color shirt am I wearing?” the then four-year-old asked. 

Danovitch, a researcher at the University of Louisville, says he was testing the boundaries of Siri’s knowledge—something that, her research finds, often happens when kids get to be about that age. And the more studies she and others in the field conduct, the more robust the behavior appears.

In one, published in May, Danovitch and two colleagues conducted a “selective trust study” on Chinese kids where five- to eight-year-olds were separated into groups and asked questions like “How many days does it take Mars to revolve around the sun?”

Danovitch and her colleagues offered a few contrasting narratives to these kids: The internet said 600 days; their teacher said 700 days. Who did they trust? (The answer, by the way, is 687 days.)

Turns out kids overwhelmingly trust a teacher—even if the teacher is wrong. That makes sense: they know their teacher, and that teacher has developed a strong relationship with them. But the kids preferred their peers over the internet, too, even though they knew their friends had roughly the same amount of knowledge as they did.

Danovitch’s theory as to why kids behave this way is that the idea of voice assistants—and by extension, the internet—is amorphous and hard to grasp. If you’re a child who thinks there's a tiny woman who lives in the kitchen called Alexa (as Danovitch says her son did), you’re trying to wrap your head not only around how this thing works but what its knowledge base is in the first place. Trusting another person, on the other hand, is hardwired into our brains.

Earlier this year, Silvia Lovato of Northwestern University presented research on how American kids in the same age range Danovitch studied are not only skeptical of voice assistants, but remarkably creative in their attempts to test how reliable the gadgets are.

Lovato says the kids would pepper the assistants with questions. Fantasy creatures were a mainstay—Lovato’s paper is titled “Hey Google, do unicorns exist?” Voice assistants are often programmed to answer “I don’t know” to these sorts of queries (Santa, Easter bunny, and the tooth fairy, etc.), making them appear less trustworthy to children.

Danovitch and Lovato’s work suggests not only that kids are way more sophisticated about technology than we think but also that we, as humans, have an ingrained sense of skepticism about unknown sources that somehow gets fuzzier as we grow older. The onslaught of fake news and the rampant disinformation campaigns now common on social media might make it seem as if we dumb humans don’t investigate sources as deeply as we should. These studies indicate otherwise: technology isn’t something we trust naturally, at least when we’re young.

“Kids are paying attention,” Danovitch says. “They’re keeping track of who knows what they’re talking about and who doesn’t. Kids don’t just blindly believe every answer they get. And we look at the internet or computer programs; they don’t blindly believe those either.”