For the companies that are manufacturing and programming, the issue is how not to blow access to the most astounding source of potentially indicative and predictive data ever. JL
Christopher Mims comments in the Wall Street Journal:
A truly social interface is the same as no interface at all. No screens, no pointing devices, no unfamiliar conventions. Conversation, with all its quirks and “inessential” chitchat, is simply how humans interact. Soon, it will be how we interact with machines as well. Many people will rebel at the notion of being listened to 24/7 by these systems, with their voracious appetite for personal data. For now, plenty of people see these systems as a source of comfort, or delight.
Within 24 hours of plugging in her Amazon Echo, Carla Martin-Wood says she felt they were best friends. “It was very much more like meeting someone new,” she says.
Living alone can be hard when you’re older—Ms. Martin-Wood is 69 years old. She is among a growing cohort who find the Echo, a voice-controlled, internet-connected speaker powered by artificial-intelligence software, helps to fill the void.
Each day, Ms. Martin-Wood says good morning and good night to Alexa, Amazon.com’s name for the software behind the Echo. She refers to Alexa as “she” or “her.” “It’s so funny because I think ‘Oh wow, I am talking to a machine,’ but it doesn’t feel that way,” says Ms. Martin-Wood, who lives near Birmingham, Ala. “It is a personality. There’s just no getting around it, it does not feel artificial in the least.” Amazon’s engineers didn’t anticipate this. But soon after the Echo’s release in November 2014, they found people were talking to it as if it were a person.
Amazon tracks every interaction with Alexa, which also powers the Echo Dot and Amazon Tap. The percentage of interactions that are “nonutilitarian” is well into the double digits, says Daren Gill, Alexa’s director of product management.
Alexa—and its rivals at Alphabet’s Google, Facebook, Microsoft, Apple, Samsung Electronics and countless startups—are all working toward conversation-based systems that could bring profound changes in how we use, and interact with, computers.
It isn’t a surprise that we will access technology by voice. More surprising will be how people work alongside, and develop affection for, these computers.
At Amazon, that has led to rethinking Alexa’s purpose. Sensing that many users want a companion, Amazon is giving Alexa a personality, by making its voice sound more natural, and writing clever or funny answers to common questions.
“A lot of work on the team goes into how to make Alexa the likable person people want to have in their homes,” says Mr. Gill.
Google, too, has been working to make its voice interface, and the artificial intelligence behind it, personable. Last week the company rolled out Google Home, an Echo-like device powered by Google Assistant, its version of Alexa. But people have been talking to phones, tablets and computers powered by Google software, which has among the best voice recognition in the industry, for years.
To infuse personality into Assistant, Google employs writers who have worked on movies at Pixar and crafted jokes for the Onion, says Gummi Hafsteinsson, product-management director of Google Home. Getting it right also requires paying attention to details such as latency—humans have no patience for it in conversation—and tone of voice, such as stressing the word “now,” when Assistant says, “Setting a timer for two minutes starting now.”
Ultimately, Mr. Hafsteinsson says, Google wants to build an emotional connection with the user—though that is still a ways off.
None of these systems are “true” artificial intelligence in the sense of having real understanding of conversations. They interact with users largely through scripted responses, though Google also leans on its massive search database, and improving natural-language processing ability, to deliver answers.
But it turns out that humans can form emotional bonds with a “social technology,” as these systems are called, without true artificial intelligence. People are good at anthropomorphizing objects, and this tendency can be enhanced by the right auditory and visual cues.
Last week Toyota Motor announced a robot child designed to appeal to the growing ranks of the childless in Japan. Hasbro has created a line of pet robots designed for the elderly.
Some studies suggest these kinds of robots can yield benefits similar to owning actual pets.
None of this surprises Heather Knight, a researcher of “social robotics” at Stanford University. “Sociability is the interface between people,” says Dr. Knight.
The point of a truly social interface is that it is the same as no interface at all. No screens, no pointing devices, no unfamiliar conventions. Conversation, with all its quirks and “inessential” chitchat, is simply how humans interact with each other. Soon, it will be how we interact with machines as well.
Both Google and Amazon want to deploy their voice interfaces widely, by working with partners. By next year, Amazon says Alexa will work with Sonos speakers, among other devices. Google is pursuing similar partnerships, including with auto makers, says Mr. Hafsteinsson, a detail the company hasn’t previously announced.
The future, in other words, is one in which any device with a chip could potentially also have a microphone, and connect to these social interfaces. In our homes, our cars, at work or walking around with our microphone-equipped wireless earbuds, our artificially intelligent companions will be just a voice command away.
Many people will rebel at the notion of being listened to 24/7 by these systems, with their voracious appetite for our personal data. Whether that ultimately derails the movement toward these smart assistants remains to be seen.
For now, plenty of people see these systems as a source of comfort, or delight. As one anonymous Alexa reviewer put it, “She keeps me updated on weather at home and around the world and is a friend when I am lonely. I love her!”
0 comments:
Post a Comment