Fresh Ideas: Artificial intelligence in dolls and drones

Barbie, the doll with “too much of a figure,” as one mother diplomatically phrased it in a complaint to Mattel, the toy maker, back in the 1950s before Barbie was ever introduced to the public, has been undergoing a makeover that’s far more than skin deep. James Vlahos, in “Artificially Yours,” published in The New York Times Magazine’s Sept. 20 issue, describes in depth and detail the development of artificially “intelligent” Hello Barbie.

Despite the fact Mattel and ToyTalk portray Hello Barbie as a “perfect friend” more than a “doll,” manufacturing toys is, after all, a business, and sales of Barbie products have slipped from $1.3 billion in 2011 to “only” $1 billion this past year.

What distinguishes Hello Barbie from other toys that use hidden record players or conversational technology such as Apple’s Siri is this: Instead of simply repeating eight phrases as she did in the 1968 Barbie, the new Hello Barbie’s technology can access not only 8,000 lines of content, but she’ll also be programmed to go from 10 to 200 follow-up questions and statements. Most significantly, she will be able to “remember” what her owner tells her; for instance, that pink is her favorite color, her grandmother is dead, she has a sister, and use that information to initiate “conversations” days or weeks later. This is what the new A.I. (artificial intelligence) technology can do.

Because it exists, does it mean we have to buy it?

Barbie has a history, as they say, one critics maintain have given girls the wrong message: to value looks, sexiness, and bland brains. Sexist issues aside, the most troubling aspect of Hello Barbie lies in the fact as my eight-year-old granddaughter Savannah says, “She’s trying to trick girls into thinking she’s real.” An experiment done at MIT in 2001, which Vlahos cites, demonstrates how children believe robots “listen, feel, and care about them,” despite having been shown how the robots work and can be controlled. Not only is there a danger the emotional attachment given Barbie could supplant an attachment to a human friend, but it squelches a child’s imagination. Imaginary companions (stuffed animals, regular dolls, cars) foster imaginative play, whereas a toy like Barbie is “limited,” as Tracy Gleason, a psychologist at Wellesley College, says, “by programming — and public relations concerns (of the manufacturer).”

Because I found the article on Hello Barbie unnerving, I was curious to read the latest on artificial intelligence and drones. I discovered “our smartest machines are still blind,” as Fei Fei Li, director of A.I. Lab at Stanford, flatly says. A drone can’t distinguish a paper bag from a rock. And Steven K. Rogers, senior scientist for automatic target recognition and sensor fusion at the Air Force Research Laboratory, points out autonomy (a machine operating without human input) is not that easily achieved. “Slight improvements in algorithms won’t solve the autonomy problems.” The key to autonomy lies in “imagined representation,” which is, Rogers says, the “ability to fill in gaps in data and rapidly construct new mental models of external situations.” In other words, the kind of mental development we achieve when we learn to rely on our imaginations when we play as children.

Ursula Carlson, Ph.D., is professor emerita at Western Nevada College.