Dec. 2, 2011
BOSTON — Think of how reliant we are on technology. You may wake to the morning alarm on your smartphone, then check email on the same phone at breakfast. In the car, you’re helped by GPS.
We are so engaged with automated systems that we don’t think about it — until something goes wrong. Look around work when the internet goes down and witness frustration, anger and sudden desperation.
So we pick up the phone for help. And we’re greeted with… yes…an automated voice.
If you’ve called a customer helpline in the past decade, chances are you’ve spoken with a non-human voice created by Nuance, the largest independent vendor ofspeech recognition technology. The company is based in Burlington, Mass.
The tone of that automated voice matters, said Peter Mahoney, Nuance’s chief marketing officer. “There’s a lot of tuning you can do to the voice. Sometimes you need to be a little bit more quiet and helpful, sometimes you need to be cheery,” he said.
“We’ll test the voices implemented in the system to see how people react. And you can really tune these things up and down based on the scenario that you’re in.”
Now, Nuance knows we just want to press zero and speak to a live person. So the company has turned speech synthesis into an art, devoting hours of research and testing to create subtleties of speech and tone in the voice systems… so we don’t react negatively.
And that is a problem, according to MIT professor of technology and society Sherry Turkle. She’s the author of the book “Alone Together,” a skeptical take on our lives in the digital realm.
“Sophistication in voice recognition and the amount of energy that’s going into the fluidity and the nuance and the tone of artificial speech is to make us more comfortable with a world where we will be happy to have artificial intelligence as first as our tellers and sales assistants and finally as companions,” Turkle said.
As big a leap that may seem from yelling at some voice recognition technology, Turkle thought it’s important to stop ourselves and ask why we demand so much from technology.
Think about areas such as caretaking for the elderly, and nannies for child rearing. Do we want robots to do these personal things?
“Why do we want to be so comfortable having robots in this job — in areas where people are the only ones who really should be there?” Turkle said.
Phil Ridarelli is the real but recorded voice behind many customer service lines, from tech support for a multinational computer company to ordering a pizza.
Ridarelli works for Interactions, a company based in Franklin, Mass. Interactions adds real people to the mix, but you never speak with them. Instead, the company’s “intent analysts” listen in on your phone calls and they send the correct automated response as soon as they figure out what you’re calling about.
The responses sound so realistic that when callers ask, “Are you real?” there’s a recording of Ridarelli saying “No, actually, I’m a pre-recorded automated response system designed to help you with all your needs.”
When the real Ridarelli recorded these prompts, he pictured himself sitting in a call center.
“We want it to sound like a character, a person. So that when you call up, you can actually imagine the individual who is sitting on the other end of the line, helping you out,” he said.
That’s a lot of time and energy for fake conversations and predetermined empathy.
These companies that create voice recognition systems try to make these interactions simpler for their clients’ customers — that means us — while keeping costs down. But Turkle thinks some things aren’t meant to be simplified.
“Human beings are trained to know a conversation when they’re in one. And my fear is that we’re losing that sense,” she said. “Conversations are complicated. They should not be simplified.”
We may be too sensible to let a robot raise a child, but the next time you ask your smartphone a question, or follow your GPS’s instructions, think about how much faith we put into a system. Because we don’t want to end up like Dave in the movie “2001: A Space Odyssey,” permanently silenced by a robot.
“Dave, this conversation can serve no purpose anymore. Goodbye.”
Comment on This Article
Subscribe to WGBH News Emails