July 28, 2014
What the Ultimate Virtual Assistant Might Look Like in the Future
With its original story of a lonely man who falls in love with an operating system, writer-director Spike Jonze’s romantic comedy Her has delighted audiences and critics alike. But the Zeitgeist-capturing premise at the film’s core is an evolution of a sci-fi staple—the computer as ultimate personal assistant.
What will the ultimate personal assistant look like? Could Her happen to you?
As much as we love our devices, it’s doubtful anyone has actually fallen in love with Apple’s Siri or Google Now. But with their ability to perform and anticipate myriad tasks through natural language recognition, these applications are the early incarnations of far more intelligent—and perhaps more lovable—personal assistants to come.
As for the specific technologies that will make up the ultimate personal assistant, the jury is still out. Will the “eyes” of the system be a single camera or multiple cameras? Will its “ears” be a single microphone or multiple mics? How will it learn to seamlessly answer, converse and anticipate—which Google has identified as the three primary functions of a search engine?
Some versions of the ultimate personal assistant suggest a larger integration with the Internet of Everything—a main focus of last month’s International Consumer Electronics Show in Las Vegas. And what are the privacy, safety and security implications of such a powerful system? Would you want, for example, an all-seeing “eye” following your every move—like Stanley Kubrick’s lip-reading HAL 9000 computer in 2001: A Space Odyssey?
A High-Tech Holy Grail
In Her, Jonze’s vision of the ultimate personal assistant uses a deceptively simple blend of technologies. Joaquim Phoenix’s love-smitten Theodore Twombly converses with his husky-voiced operating system, Samantha (Scarlett Johansson), via a removable earpiece and either a desktop computer or, when he’s on the move, a wallet-sized mobile device.
But Samantha’s virtuosic conversational skills and emotional intelligence are light years ahead of today’s technologies. In some scenes, her abilities would require massively scaled real-time image recognition, spatial understanding, facial and mood recognition, and a grasp of social cues, according to Dag Kittlaus, co-founder and former CEO of Siri, which he sold to Apple in 2010.
In a recent article in the entertainment-trade magazine “Variety,” Kittlaus writes that when he and his co-founders created Siri in 2007, they were “trying to build the world’s first true virtual personal assistant, to make interacting with your devices as simple as a conversation.” Their efforts were “seriously influenced” by the HAL 9000 as well as the talking computers from Star Trek and KITT from the “Knight Rider” TV series.
“Siri was built to get things done,” Kittlaus writes. “But then a funny thing happened. Siri blew up into a cultural phenomenon overnight. It wasn’t just a new and easier way to use your phone. Siri was fun. It felt a little human. Millions of people would chat with Siri for hours at a time.”
Google ‘s engineering director, Scott Huffman, recently stated that the company’s goal is to create the ultimate personal assistant. Interaction with it will not depend on a computer screen, Huffman said, but will take the form of back-and-forth conversations—in the car or living room, for instance. These interactions may occur via a large screen or via microphones and speakers that wait for the user’s verbal prompt.
If the idea of a living space fitted with networked mics leaves you a tad queasy, you’re not the only one. Digital technology policy researcher Adam Thierer says he finds the potential security concerns worrisome. But he’s no fan of squelching creativity either, favoring what he calls “permissionless innovation” over the “precautionary principle.” (His new book, “Permissionless Innovation: the Continuing Case for Comprehensive Technology Freedom,” is set to debut this month.)
A senior research fellow at the Mercatus Center at George Mason University in Fairfax, Va., Thierer says the ultimate personal assistant will bring its share of trade-offs—like other technologies from Facebook to Google Glass to the automobile.
“But I hope we don’t rush to pre-empt these things in a top-down fashion,” he says. “In time, I’d probably integrate it into my life.”
Either way, Google is moving toward its stated goal. With the help of its Knowledge Graph—a knowledge base used to enhance search results—search is already evolving away from keywords and the “10 blue links” that became the norm after Google launched its search tool 15 years ago.
Last year, the company launched a new “Hummingbird” algorithm designed to facilitate a more human way to interact with users and provide more direct answers. It also introduced the term “hot-wording” to describe search using the verbal prompt “OK Google,” without the need for an interface.
Next up, Huffman acknowledged in December, Google may open the Knowledge Graph to third-party developers. That would bring the dream of the ultimate personal assistant even closer—while magnifying potential privacy and other concerns.
“But if we spend all our time obsessing over and trying to plan public policies around hypothetical worst-case scenarios, then the best-case scenarios will never come about,” Thierer says.
And falling in love with a virtual personal assistant? Could best-case scenarios (or worst-case ones, depending on your stance) include that? The jury’s out on that, too.
Laurence Cruz is a freelance writer based in Los Angeles. A U.K. transplant, he has worked as a reporter with The Associated Press in Seattle and as an environmental reporter for The Statesman Journal in Salem, Oregon. He has a BA in English from Oxford and an MA in Communications from Washington State University. Used with the permission of http://thenetwork.cisco.com/.