On December 10th, TIME Magazine’s website published a story by Zeke Miller and Denver Nicks titled “Meet the Robot Telemarketer Who Denies She’s A Robot.”
It all began when TIME Washington Bureau Chief Michael Scherer got a phone call from a pleasant woman offering a deal on health insurance. Scherer sniffed out that something wasn’t right, and when he asked the caller “point blank if she was a real person, or a computer-operated robot voice, she replied enthusiastically that she was real, with a charming laugh. But then she failed several other tests. When asked ‘What vegetable is found in tomato soup?’ she said she did not understand the question. When asked multiple times what day of the week it was yesterday, she complained repeatedly of a bad connection.”
TIME reporters got busy trying to verify who – or what – the caller was. As a reporter relates it: “Her name, she said, was Samantha West, and she was definitely a robot, given the pitch perfect repetition of her answers. Her goal was to ask a series of questions about health coverage –‘Are you on Medicare?’ etc. – and then transfer the potential customer to a real person, who could close the sale.”
You can listen to a few of the calls here: https://newsfeed.time.com/2013/12/10/meet-the-robot-telemarketer-who-denies-shes-a-robot. The caller has a pleasant voice, and yes, the laugh is charming. And the script is pretty good.
In a subsequent story published December 17, Denver Nicks gives us Part 2: It turns out that Samantha West “functions much like a remote-controlled car, directly operated by a real person working in a call center outside the United States.” A representative of the company that deploys Samantha said “she” is a computer program that lets telemarketers outside the U.S., who may have non-American accents, qualify leads and send good prospects to native speakers in America. The rep explained:
“When Samantha West calls, there is a person on the other end of the line … who is an active participant in the conversation. That the person was limited to communicating through a machine with pre-recorded utterances did not change the fact … that there was another human participating in the conversation.”
What readers thought
The comments following the posts ranged from rants against robocalls to a spirited discussion of whether tomatoes are vegetables or fruits, to “The robots are sentient now, but the only jobs they can get are in telemarketing. One commenter’s thoughts sound informed:
“I have to agree with (name). This is too smooth for synthesized voice and too coherent for automated scripts. It is likely prerecorded responses punched up by an attendant. This allows a smoother pitch and cheaper labor, than for ‘live voice’ attendants. So Samantha was not lying. ‘She’ is a human choosing each response, and ‘her’ voice is/was human. Though the script is well read, and well managed, it is inherently inauthentic. It also is necessarily limited in response options and vocabulary.”
Renaissance man Ray Kurzweil is a director of engineering at Google, and an expert on text-to-speech synthesis, artificial intelligence, and electronic keyboard instruments (among many, many, many other things). He’s probably the world’s foremost champion of the “singularity” – which Slate describes as a point in human progress at which our machines become as smart as we are. When we get there (assuming we do), will those machines be calling to sell us something? And will they be smart enough to contact us only about the things we really want to buy? Will they actually be drones operated by live persons?
It’s probably a great savings to use semi-automated robocalls such as the Samantha program, but I wonder: is the trade-off for authenticity worth it? And what happens when these programs get so sophisticated that we can’t tell that they aren’t real people speaking in real time?
What do YOU think?