“Your domestic problems are completely solved.” So says a robotics technician to a grateful housewife in “Leave It to Roll-Oh,” a promotional film produced by the Chevrolet Motor Company for the 1939 New York World’s Fair. The titular star of the picture, a “chromium-plated butler,” is an ambulatory automaton that looks like a beefy version of the tin man from The Wizard of Oz. Operated by remote control, the contraption can be commanded to perform various chores at the push of a button: Clean House, Get Dinner, Wash Dishes, Fix Furnace.
Although “just a daydream,” as the movie’s narrator comes to reveal, Roll-Oh personified the common conception of a household robot. From the moment we first imagined having mechanical servants at our beck and call, we’ve assumed they would be constructed in our own image. Outfitted with arms and legs, heads and torsos, they would perform everyday tasks that we’d otherwise have to do ourselves. From The Jetsons’ indefatigable maid Rosie, to the officious droid C-3PO in Star Wars, to Westworld’s tortured “host” Dolores Abernathy, the robotic helpmates of popular culture have been humanoid in form and function.
I’s time to revise our assumptions. A robot invasion of our homes is under way, but the machines — so-called smart speakers like Amazon Echo, Google Home, and the forthcoming Apple HomePod — look nothing like what we anticipated. Small, squat, and stationary, they resemble vases or cat-food tins more than they do people. Echo and its ilk do, however, share an important trait with their imaginary forebears: They illuminate the times. Whatever their shape, robots tell us something vital about our technologies and ourselves.
Smart speakers have been around just three years, but they already have a hold on us. Powered by “chatbots” like Siri and Alexa, the devices are in the midst of a sales boom. Some 35 million Americans now use the diminutive, talking computers — more than twice the number of just a year ago, according to estimates by eMarketer — and analysts predict sales will continue to surge in the coming months. Google just expanded its Home line, and Microsoft, Samsung, Facebook, and China’s Alibaba are all expected to enter the market soon.
The allure of the gadgets is obvious. Smart speakers are oracles of the countertop. They may not speak for the gods, but they do deliver useful reports on news, traffic, and weather. And they have other talents that their Delphic ancestor couldn’t even dream of. They can serve as DJs, spinning playlists of everything from blue-eyed soul to British grime. They can diagnose ailments and soothe anxieties. They can summon taxis and order pizzas. They can read bedtime stories to toddlers. They can even bark like a watchdog to scare off burglars. And they promise to be the major-domos of home automation, adjusting lights and thermostats, controlling appliances, and issuing orders to specialized robots like the Roomba vacuum cleaner.
Still, if you were looking forward to having a Rosie scurrying around your abode, feather duster in hand, an Echo feels like a letdown. It just sits there.
There are good reasons the domestic robot has taken such an uninspiring shape. Visualizing a nimble, sure-footed android is easy, but building one is hard. As Carnegie Mellon professor Illah Nourbakhsh explains in his book Robot Futures, it requires advances not only in artificial intelligence but in the complex hardware systems required for movement, perception, and dexterity. The human nervous system is a marvel of physical control, able to sense and respond fluidly to an ever-changing environment. Just maintaining one’s balance when standing upright entails a symphony of neural signals and musculoskeletal adjustments, almost all of which take place outside conscious awareness.
Achieving that kind of agility with silicon and steel lies well beyond the technical reach of today’s engineers. Despite steady progress in all fields of robotics, even the most advanced of today’s automatons still look and behave like parodies of human beings. They get flustered by mundane tasks like loading a dishwasher or dusting a shelf of knickknacks, never mind cooking a meal or repairing a furnace. As for multitalented robots able to shift flexibly among an array of everyday tasks: they remain science-fiction fantasies. Roll-Oh is still a no-go.
Meanwhile, thanks to rapid gains in networking, natural language processing, and miniaturization, it’s become simple to manufacture small, cheap computers that can understand basic questions and commands, gather and synthesize information from online databanks, and control other electronics. The technology industry has enormous incentives to promote such gadgets. Now that many of the biggest tech firms operate like media businesses, trafficking in information, they’re in a race to create new products to charm and track consumers. Smart speakers provide a powerful complement to smartphones in this regard. Equipped with sensitive microphones, they serve as in-home listening devices — benign-seeming bugs — that greatly extend the companies’ ability to monitor the habits and needs of individuals. Whenever you chat with a smart speaker, you’re disclosing valuable information about your routines and proclivities.
Beyond the technical and commercial challenges, there’s a daunting psychological barrier to constructing and selling anthropomorphic machines. No one has figured out how to bridge what computer scientists term the “uncanny valley” — the wide gap we sense between ourselves and imitations of ourselves. Because we humans are such social beings, our minds are exquisitely sensitive to the expressions, gestures, and manners of others. Any whiff of artificiality triggers revulsion. Humanoid robots seem creepy to us, and the more closely they’re designed to mimic us, the creepier they become. That puts roboticists in a bind: the more perfect their creations, the less likely we’ll want them in our homes. Lacking human features, smart speakers avoid the uncanny valley altogether.
Although they may not look like the robots we expected, smart speakers do have antecedents in our cultural fantasy life. The robot they most recall at the moment is HAL, the chattering eyeball in Stanley Kubrick’s sci-fi classic 2001: A Space Odyssey. But their current form — that of a standalone gadget — is not likely to be their ultimate form. They seem fated to shed their physical housing and turn into a sort of ambient digital companion. Alexa will come to resemble Samantha, the “artificially intelligent operating system” that beguiles the Joaquin Phoenix character in the movie Her. Through a network of tiny speakers, microphones, and sensors scattered around our homes, we’ll be able to converse with our solicitous AI assistants wherever and whenever we like.
Facebook founder and CEO Mark Zuckerberg spent much of last year programming a prototype of such a virtual agent. In a video released in December, he gave a demo of the system. Walking around his Silicon Valley home, he conducted a running dialogue with his omnipresent chatbot, calling on it to supply him with a clean t-shirt and toast bread for his breakfast, play movies and music, and entertain his infant daughter Max in her crib. Hooked up to outside cameras with facial-recognition software, the digitized Jeeves also acted as a sentry for the Zuckerberg compound, screening visitors and unlocking the gate.
Whether real or fictional, robots hold a mirror up to society. If Rosie and Roll-Oh embodied a twentieth-century yearning for domestic order and familial bliss, smart speakers symbolize our own, more self-absorbed time.
It seems apt that, as we come to live more of our lives virtually, through social networks and other simulations, our robots should take the form of disembodied avatars dedicated to keeping us comfortable in our media cocoons. Even as they spy on us, the gadgets offer sanctuary from the unruliness of reality, with its frictions and strains. They place us in an artificial world meticulously arranged to suit our bents and biases, a world that understands us and shapes itself automatically to our desires. Amazon’s decision to draw on classical mythology in naming its smart speaker was a masterstroke. Every Narcissus deserves an Echo.
This essay appeared originally, in a slightly shorter form and under the headline “These Are Not the Robots We Were Promised,” in the New York Times.